**About the Special Issue Editors**

**Mari ´an Lehocky´** is currently an Associate Professor in Physical Chemistry. After receiving his Masters degree in Physical and Applied Chemistry at the Faculty of Chemistry, Brno University of Technology, he moved on to a PhD in Chemistry and Technology of Macromolecular Substances at Tomas Bata University in Zl´ın. He spent time at the University of Aveiro, Portugal, as an Marie Curie Fellow, and he is currently a senior researcher at the Centre of Polymer Systems at Tomas Bata University in Zlin and a director of the Department of Lipids, Detergents and Cosmetics Technology at Faculty of Technology, Tomas Bata University in Zl´ın. His current scientific interests are polymer surface modification, emulsion and suspension phenomena, surface and interface science, colloid phenomena, antibacterial and atithrombotic surface modification and interaction of materials with cell systems and tissues, including STEM cells. He is an author or co-author of 80 original per-reviewed manuscripts in prestigious scientific journals (h-index 21) and inventor of 8 national and European patents. He is a member of Editorial board of Polymers, Materials and Design, and Marerials Science in Semiconductor Processing. In 2014, he won the "Molecules Best Paper Award" (2nd prize).

**Petr Humpol´ıˇcek** is interested in biomaterials, genetics and interaction of cells and materials. He is actually a leader of the research group Bioactive polymer systems at research unit The Centre of Polymer Systems of Tomas Bata University in Zl´ın. His research interests centre around the evaluation of biocompatibility of materials, especially conducting polymers. Special attention pay to the preparation of biomimetic materials, structured surfaces and development of cultivation techniques mimicking the in vivo environment. Petr Humpol´ıcek is author or co-author of 82 original ˇ peer-reviewed manuscripts in prestigious scientific journals (h-index 16) and the inventor of 6 national and European patents. He was or is the main investigator of four grants funded by the Czech Science Foundation.

### *Editorial* **Polymer Biointerfaces**

#### **Marián Lehocký 1,2,\* and Petr Humpolícek ˇ 1,2**


Received: 24 March 2020; Accepted: 30 March 2020; Published: 2 April 2020

Polymer biointerfaces are considered suitable materials for the improvement and development of numerous applications. The optimization of polymers' surface properties can control several biological processes, such as cell adhesion, proliferation, viability, and enhanced extracellular matrix secretion at biointerfaces [1–8].

Various routes of polymer biointerfaces preparation are targeted for numerous applications in the biomedical, biochemical, biophysical, biotechnological, food, pharmaceutical, and cosmetic fields [9–18].

This Special Issue, which consists of 18 articles written by research experts, reports on the most recent research on polymer biointerfaces. Several novel advanced methods related to polymer biointerfaces preparation, modification, analysis, and characterization are introduced.

Firstly, Ozaltin and co-workers treated a polyethylene terephthalate (PET) surface, for use in blood-contacting devices, by DC air plasma and immobilized fucoidan from *Fucus vesiculosus* (FU) on it, at different pH values in the range of 3–7. FU immobilization onto the PET surface after plasma treatment was found to be optimal at pH 5, as supported by FTIR, SEM, and XPS results, and provided the highest anticoagulant activity, more than 100 s, which indicates that the resulting FU-immobilized sample is an efficient anticoagulant, suitable for blood-contacting PET devices. Surface characterization was carried out by a wettability test, scanning electron microscopy, X-ray photoelectron spectroscopy, and Fourier-transform infrared spectroscopy. The anticoagulation activity of the samples was determined on the basis of prothrombin time, activated partial thromboplastin time, and thrombin time [19].

Smilek et al. prepared a hydrogel from oppositely charged biopolymer polyelectrolyte and surfactant in micellar form, at a surfactant/biopolymer charge ratio at least equal to one. Hyaluronan acted as a negatively charged biopolymer, whereas DEAED (amino-modified dextran) was used as a positively charged biopolymer. The former interacted with Septonex (carbethopendecinium bromide), whereas the latter interacted with sodium dodecylsulphate. The rheological properties of hyaluronan-based hydrogels were mainly dependent on the polymer molecular weight. Surfactant concentration (more precisely, the concentration of micelles at a surfactant/biopolymer charge ratio above 1) showed only a small effect. Surfactant concentration was found to have a much greater effect for DEAED-based hydrogels. The authors focused in more detail on the effects of various processing parameters on the properties of similar gels prepared using a slightly different cationic surfactant (approved for use in pharmaceutical formulations) and also investigated a reversely charged system—a positively charged polyelectrolyte (cationized dextran) and an anionic surfactant [20].

Mokrejš and co-workers introduced the preparation of gelatines from by-product collagen raw materials derived from the slaughter of chicken (chicken feet). Gelatine is a water-soluble protein that is obtained from collagenous raw materials by partial hydrolysis. The technological innovation consists in the biotechnological processing of (purified) feedstock by a commercial food endoprotease, which, in contrast to acidic (type A gelatines) or alkaline (type B gelatines) processing, has a variety of economic, technological, and environmental advantages [21].

Belka et al. contributed with current investigations employing MG63 cells grown on Fe-MEPE (metallo-supramolecular coordination polyelectrolyte) modified substrates, which suggest initiation of osteogenic differentiation by both high cell activity and altered morphology of the cells and/or cluster formation. The obtained results led to the conclusion that these surfaces individually support the specification of cell differentiation toward lineages that correspond to the natural commitment of the particular cell types. The authors, therefore, propose that Fe-MEPEs may be used as a scaffold for the treatment of defects in muscular or bone tissues [22].

Karakurt and collaborators investigated the antibacterial activity and cytotoxicity of immobilized glucosamine (GlcN)/chondroitin sulfate (ChS) on polylactic acid. The antibacterial surface modification of polylactic acid films was achieved through the immobilization of GlcN and ChS on film surfaces via plasma treatment, followed by acrylic acid grafting. It was found that the developed GlcN/ChS-coated polylactic acid films are excellent bactericide agents against representative Gram-positive and Gram-negative bacteria. Plasma-treated films immobilized with ChS and GlcN, separately and in combination, demonstrated bactericidal effect against strains of both bacterial types, and the results also revealed the absence of synergistic effect on antibacterial action for the combination. [23].

Lee and co-workers successfully coated Ti substrates with polycaffeic acid and metallic silver using a facile UV light-assisted method for implant applications. They confirmed that the coating process was successful by SEM and AFM analyses. At the same time, they verified the deposition of polycaffeic acid and metallic silver by confirming the elemental composition through XPS, EDS, and mapping methods, and the physical properties (hydrophilicity) of the samples were verified using water contact angle measurements. In vitro biocompatibility and antibacterial studies showed that polycaffeic acid with metallic silver can inhibit bacterial growth, while proliferation of MC-3T3 cells was observed. Therefore, the obtained results suggest that the introduced approach can be considered as a potential method for functional implant coating in the orthopedic field [24].

Bernal-Ballen et al. examined the development of a bioartificial polymeric material made of polyvinyl alcohol (PVA), chitosan (CHI), and fucoidan (FUC) and the incorporation of ampicillin as an antibacterial agent. The prepared films were tested, and it was elucidated that the bioartificial polymeric material has potential for inducing cell regeneration in vitro. The characterization techniques used in the manuscript indicated that PVA brings water resistibility to the system, whereas CHI and FUC are responsible for creating a porous microstructure, which allows the cells to adhere to and grow within the matrix. The obtained information indicated that PVA, CHI, and FUC are compatible, as evidenced by FTIR spectra and SEM images. The new material is an outstanding candidate for cell regeneration, as a result of the synergic effect that each component provides to the blend. [25].

Vitkova et al. obtained nanofibers containing hyaluronic acid (HA) by solution electrospinning. Two approaches were chosen: co-electrospinning of aqueous blend solutions of hyaluronic acid/polyvinyl alcohol and hyaluronic acid/polyethylene oxide and use of an intermediate solvent for electrospinning of pure hyaluronic acid solutions. The choice of materials was done with regard to their potential utilization for cell cultivation. The influences of polymer concentration, average molecular weight (*M*w), viscosity, and solution surface tension were analyzed. HA and PVA were fluorescently labeled in order to examine the electrospun structures using fluorescence confocal microscopy [26].

Dvoˇráková and co-workers demonstrated a new hydrophilization technique based on plasma deposition of a thin film using mixtures of propane/butane with nitrogen at atmospheric pressure. Unlike a simple plasma treatment, the observed high-surface free energy values were due to the properties of the deposited plasma–polymer nanolayer. Therefore, the wettability improvement did not depend on the substrate material, and the aging of the surface modification was highly reduced. The thin layers of the prepared plasma–polymer exhibit highly stable wetting properties, are smooth, homogeneous, and flexible, and adhere well to the surface of polypropylene substrates. Moreover, they are constituted of essential elements only (C, H, N, O). This makes the presented modified plasma–polymer surfaces interesting for further studies in biological and/or technical applications [27].

Habib et al. grafted ascorbic acid onto a polyethylene surface via plasma treatment in order to improve its antimicrobial effects. Plasma treatment was effectively used as a radical initiator with subsequent incorporation of ascorbic acid, which served as an antimicrobial agent, on the polyethylene surface. This modification was confirmed by the enhanced wettability and adhesion properties. The results showed changes in the wettability, adhesion, and roughness of the polyethylene surface after plasma treatment as well as after ascorbic acid grafting. This is a positive indication of the possibility of grafting ascorbic acid onto polymeric materials using plasma pretreatment, enhancing its antibacterial activity [28].

Skopalová and co-workers studied polyaniline films modified by substances with anticipated anticoagulant activity, sodium dodecylbenzenesulfonate, 2-aminoethane-1-sulfonic acid, and *N*-(2-acetamido)-2-aminoethanesulfonic acid. The hemocompatibility tests conducted on these polyaniline films confirmed the absence of anticoagulation activity, though the functional groups typical of anticoagulation substances were present. Hemocompatibility is an essential prerequisite for the application of materials in the field of biomedicine and biosensing. In addition, mixed ionic and electronic conductivity of conducting polymers is an advantageous property for these applications. The results showed that the anticoagulation activity was highly affected by the presence of suitable functional groups originating from the used heparin-like substances and by the properties of the polyaniline polymer itself [29].

Urbánková et al. used sodium caseinate in order to stabilize emulsions containing bioactive tamanu and black cumin oils. The emulsions were prepared by ultrasound treatment or high-shear homogenization with Ultra-Turrax. The analysis of the oils' fatty acid composition revealed a higher degree of unsaturation for cumin oil, with higher content of linoleic acid C18:2, which corresponded to the higher iodine value determined for this oil. The antibacterial activities of both oils and of their emulsions were investigated with respect to the growth suppression of common spoilage bacteria, using the disk diffusion method. The oils and selected emulsions were proven to act against Gram-positive strains, mainly against *Staphylococcus aureus* and *Bacillus cereus*. Regrettably, Gram-negative species were fully resistant to their action [30].

St'ahel and co-workers deposited oxazoline-based thin films on glass substrates using atmospheric pressure dielectric barrier discharge. This study presents a new way to produce plasma-polymerized oxazoline polymers, which are a new promising class of polymers for biomedical applications, with antibiofouling properties and good biocompatibility. The authors describe the film preparation procedure. Nitrogen was used as the working gas for the discharge, 2-methyl-2-oxazoline vapors, used as the monomer, were admixed to the nitrogen flow. This gas composition made it possible to obtain a homogeneous discharge, which led to the deposition of homogeneous thin films. To improve the film properties, it was necessary to increase the substrate temperature during the deposition. All deposited films are cytocompatible and exhibited excellent antibacterial properties against *S. epidermidis*, *S. aureus*, and *Escherichia coli* [31].

Shah et al. developed a novel dual crosslinked film for promising future applications in ophthalmology, skin tissue engineering, and wound dressing. Firstly, a collagen/chitosan film was prepared by the solvent casting technique, utilizing two crosslinking agents together, i.e., tannic acid and genipin. The obtained final dual crosslinked film was translucent, thin, and greenish-blue in color. Enzymatic degradation of the films was performed with lysozyme and lipase. Cell adhesion and proliferation were tested using mouse embryonic cell lines, by culturing the cells on the dual crosslinked film. These dual crosslinked polymeric film find their application in ophthalmology, especially as implants for temporary injured cornea and skin tissue regeneration [32].

Vesel and co-workers presented results regarding systematic 2D mapping of surface wettability, which provide an insight in processes responsible for surface activation of the polymer polyethylene terephthalate using a simple atmospheric-pressure plasma jet. The discharge tube was only flushed with Ar gas to get rid of permanent gases, such as nitrogen, oxygen and CO2, but the water remained on the surfaces, due to the humidity and laboratory atmosphere. In the case of a rapid activation, a

very sharp interphase between the activated and the unaffected surface was observed and explained by the peculiarities of high-impedance discharges sustained in Ar with the presence of impurities from water vapor. The results obtained by X-ray photoelectron spectroscopy confirmed that the activation was a consequence of functionalization with oxygen functional groups [33].

Narimisa et al. applied a plasma source to deposit thin layers under atmospheric pressure. Using different techniques, the effect of process parameters such as applied power, carrier gas flow rate, distance from capillary to the substrate, and treatment time on the deposition efficiency was studied. A high level of monomer fragmentation observed with optical emission spectroscopy, together with the non-uniform distribution of the monomer observed by computational fluid dynamic simulations, was shown to be a reliable indicator of coating quality. By following this research strategy, crucial information regarding the effectiveness of atmospheric-pressure microwave plasma jet for thin film deposition can be revealed [34].

Villamil Ballesteros and co-workers decellularized bovine amniotic membranes using four different protocols, and the differences in terms of decellularization were considered as negligible. All membranes provided DNA concentrations <50 ng/mg, indicating that traces of the nucleic acid were present in the prepared material, although the obtained values were negligible, which implies that the decellularized membranes did not contain native cells from the bovine amniotic membranes. The in vitro biocompatibility of the studied samples was demonstrated. Hence, these matrices may be deemed as a potential scaffold for epithelial tissue regeneration [35].

Finally, Kluˇcáková studied the influence of humic acids on the transport of metal ions and dyes in agarose hydrogels. It was confirmed that humic acids retarded the transport of diffusion probes. Humic acids' enrichment caused decreases in the values of effective diffusion coefficients, due to their complexation with the diffusion probes. The effect of complexation was selective for a particular diffusion probe. The aim of this study was to investigate the influence of the interactions between humic acids and probes on hydrogels' release ability, as hydrogels play an important role in the monitoring of the mobility of pollutants in nature as well as in their removal and in water treatment. They are usually based on materials able to absorb water and different pollutants in their structure [36].

**Acknowledgments:** First of all, we would like to express my deep gratitude to *Polymers*, especially to its Editorial office for continuous guide, care, and help during all steps of the preparation and production of this Special Issue titled "Polymer Biointerfaces". We would also like to extend our gratitude to all contributing authors for their valuable manuscripts as well as to all reviewers who have helped with valuable suggestions.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Agarose Hydrogels Enriched by Humic Acids as the Complexation Agent**

#### **Martina Kluˇcáková**

Faculty of Chemistry, Brno University of Technology, Purky ˇnova 118/464, 612 00 Brno, Czech Republic; klucakova@fch.vutbr.cz; Tel.: +42-054-114-9410

Received: 26 February 2020; Accepted: 17 March 2020; Published: 19 March 2020

**Abstract:** The transport properties of agarose hydrogels enriched by humic acids were studied. Methylene blue, rhodamine 6G and Cu(II) ions were incorporated into hydrogel as diffusion probes, and then their release into water was monitored. Cu(II) ions as well as both the dyes studied in this work have high affinity to humic substances and their interactions strongly affected their diffusion in hydrogels. It was confirmed that humic acids retarded the transport of diffusion probes. Humic acids' enrichment caused the decrease in the values of effective diffusion coefficients due to their complexation with diffusion probes. In general, the diffusion of dyes was more affected by the complexation with humic acids in comparison with Cu(II) ions. The effect of complexation was selective for the particular diffusion probe. The strongest effect was obtained for the diffusion of methylene blue. It was assumed that metal ions interacted preferentially with acidic functional groups. In contrast to Cu(II) ions, dyes can interact with acidic functional groups, and the condensed cyclic structures of the dye probes supported their interactions with the hydrophobic domains of humic substances.

**Keywords:** hydrogel; agarose; humic acid; reactivity; diffusion

#### **1. Introduction**

Hydrogels play an important role in the monitoring of the mobility of pollutants in nature as well as in their removal and water treatment. They are usually based on the materials able to absorb water and different pollutants in their structure. There are many studies and reviews dealing with bio-polymeric and polymeric hydrogels for different environmental applications. Agarose, pectin, alginate or lignin may be mentioned as examples of materials eligible for these purposes [1–9]. This work is focused on the agarose and its hydrogels as a medium for the investigation of their transport properties and effect of humic substances on the release of metal ions and dyes from them. Agarose hydrogels have a number of practical utilizations. They can be, e.g., used as separation media in column chromatography and act as bacterial culture support [1,2]. Their gelling [3], rheological and thermal properties [7–9], and internal structure [10–12] have been widely studied, but the effect of micro-scale structural factors of porous media on effective mass diffusion is not well understood [11], and the molecular structure of agarose is still a matter of debate [3,7]. Some studies stated that the hydrogels consist of thick bundles of agarose chains and large pores of water [10,13], which constitute the main paths for diffusing particles. The hydrogel is widely used as a transport medium for the determination of diffusion characteristics of different molecules and ions [2,13–16]. Pluen et al. [14] studied the diffusion of several different macromolecules through 2% agarose hydrogel. Data were analysed by means of Yimm-Rouse model [17] and the reptation model [13,18]. Gong et al. [13] investigated the effect of aspect ratio of protein on its diffusion in hydrogel and the effect of electrostatic interaction between protein and hydrogel on its transport through hydrogel. The influence of of electrostatic and specific interactions on the diffusion and partitioning of various solutes in agarose

hydrogels was studied by Fation-Rouge et al. [2]. Gutenwik et al. [15] determined diffusion coefficients of proteins at different pH values and ionic strengths by means of diffusion cells. Golmohamadi et al. [16] measured the self and mutual diffusion of different cations and correlated them with Donnan potentials of hydrogels. Wang et al. [19] characterized the diffusion of cations and anions in thin films of agarose hydrogel. Agarose hydrogel is often used as transport medium passive samplers based on diffusion gradients in thin hydrogel films [19,20].

As can be seen, agarose hydrogels are widely studied and utilized as transport media for different diffusing particles. In our previous studies [21–24], agarose hydrogels were enriched by humic acids as an active component in order to support the complexation of diffusing particles in hydrogels. Humic-agarose hydrogels were characterized by means of their viscoelastic properties [21,24], and the transport of different ions through the hydrogels was studied by means of diffusion cells [21,23,24] and non-stationary transient diffusion [22,23]. The effect of acidic functional groups of humic acids on the complexation and transport of metal and dye ions was investigated by means of the selective blocking of carboxylic groups by methylation [23,24]. Our experiments showed that the addition of humic acids in agarose hydrogel can strongly influence the complexation and diffusion of metal ions and dyes, which resulted in changes in effective diffusion coefficients.

In this study, a different type of experiment was performed. The study is focused on transport properties of agarose hydrogels enriched with humic acids, especially the release of different diffusion probes from the hydrogels. Agarose hydrogel was enriched by humic acids and also by metal or dye ions. The transport out of hydrogel was monitored and the diffusion characteristics of the transport were determined. Simultaneously, the degree of immobilization of ions was calculated, and the ratio between mobile and complexed ions was calculated. The aim was to investigate the influence of interactions of humic acids with probes on their release ability.

#### **2. Experimental**

#### *2.1. Materials*

Agarose (AG; routine use class), CuCl2·2H2O (p.a.), methylene blue hydrate (MB; CI basic blue 9) and rhodamine 6G (RH; CI basic red 1) were purchased from Sigma-Aldrich (St. Luis, MO, USA).

Samples of humic acids were purchased from the International Humic Substances Society (IHSS, St. Paul, MN, USA). Elliot soil humic acids (ESHA), Pahokee peat humic acids (PPHA), Suwannee river humic acids (SRHA) and Leonardite humic acids (LEHA) were used in this study. The main characteristics such as elemental composition and the contents and properties of acidic functional groups can be found on the website of the International Humic Substances Society (IHSS).

#### *2.2. Preparation of Hydrogels*

The preparation of hydrogels was based on the thermo-reversible gelation of AG aqueous solution. An accurately weighed amount of AG was dissolved in deionized water or in an aqueous solution of humic acids. The mixture was slowly heated with continuous stirring up to 80 ◦C and stirred at this temperature in order to obtain a transparent solution, and finally sonicated (1 min) to remove gasses. AG hydrogels were prepared using 1 wt % AG solution [21–24]. Afterwards, the AG solution was slowly poured into the polymethylmethacrylate (PMMA) spectrophotometric cuvette (inner dimensions: 10 mm × 10 mm × 45 mm). The cuvette orifice was immediately covered with pre-heated plate of glass to prevent drying and shrinking of gel. Flat surface of the boundary of resulting hydrogels was provided by wiping an excess solution away. Gentle cooling of cuvettes at the laboratory temperature led to the gradual gelation of the mixture [22,23]. AG–HA hydrogels were prepared from 1 wt % AG solution containing 0.01 wt % of HA. The AG and humic contents in final hydrogels were chosen on the basis of our previous results and experimental experiences [21–24]. Images of pure agarose hydrogel and hydrogel enriched by humic acids in cuvettes are shown in Figure S1 (from the Supplementary Materials).

Aqueous solutions of CuCl2, MB, and RH were used as the donor solutions for the incorporation of the diffusion probes into hydrogels. Their initial concentrations were equal to 0.1 mol.dm−<sup>3</sup> for Cu(II) salt and 1 mg.dm−<sup>3</sup> for dyes. The incorporation of the probes into hydrogels was based on diffusing of Cu(II) ions and dyes into the hydrogels. The cuvettes filled with hydrogels were placed into stirred donor solutions (4 cuvettes in 200 cm3). The diffusion probes have been diffusing into the hydrogel until a constant concentration throughout the whole hydrogel was achieved [25,26].

#### *2.3. Di*ff*usion-Release Experiments*

The cuvettes with AG and AG–HA hydrogels enriched by diffusion probes were placed in stirred distilled water (4 cuvettes in 200 cm3). The release of diffusion probes into water was monitored over time. The concentrations of probes in leachates were measured by means of UV-VIS spectrometer Hitachi U3900H (Hitachi, Tokyo, Japan). The data were used for the calculation of diffusion fluxes from the hydrogels into water through the square orifices of the cuvettes.

Simultaneously, the distributions of diffusion probes in hydrogels were determined in selected time intervals. The cuvettes were taken out of the leachates and the UV-VIS spectra were measured at various distances from the orifice by means of Varian Cary 50 UV-VIS spectrophotometer equipped with the special accessory providing controlled fine vertical movement of the cuvette in the spectrophotometer. Using the collected UV-VIS spectra, the concentrations of the probes were determined at different positions in gels [22]. The obtained data were used to compute the concentration profiles of probes in the cuvettes. The diffusion fluxes determined as the differences between the total contents of probes in hydrogels before diffusion experiments and the contents in hydrogels at given times should be the same as the values calculated on the basis of the concentrations measured in leachates; therefore their values were determined by two different measurements and averaged. All experiments were performed at laboratory temperature (25 ± 1 ◦C). Data are presented as average values with standard deviation bars. Schematic illustration of release experiment is shown in Figure S2 (from the Supplementary Materials).

#### **3. Results and Discussion**

In this work, the effect of standard humic acids as the complexation agents added in agarose hydrogels was studied by means of so-called diffusion-release experiments. Humic acids are known as substances which complex effectively with metal ions [25–33] and dyes [21–24]. Carboxylic functional groups, as well as aromatic structures and π−π interactions, are important in their reactivity. The amounts of diffusion probes in hydrogels differed slightly according to type of added humic acids (Table 1).


**Table 1.** Concentrations of diffusion probes in the AG and AG–HA hydrogels before diffusion-release experiments.

The contents differed more in the case of organic dyes. Their amounts in hydrogels without humic acids seem to be higher in comparison with enriched hydrogels. In contrast, the content of copper is slightly higher. It is well known that humic acids have very high affinity to Cu(II) ions [24–28,30–33]. Therefore, copper is a traditional model metal used to study humic reactivity. The increase in the content of Cu(II) ions in agarose hydrogels enriched by humic acids can be considered as the result of this humic affinity observed also in our previous studies [24–28]. If we compare the contents of Cu(II) ions in hydrogels containing different humic acids with the contents of their acidic functional groups declared by IHSS [34,35], we can find that the content of Cu(II) ions in hydrogels increases with the

increasing total acidity of studied humic acids. This confirmed that the acidic functional groups play the most important role in the interactions of humic acids with metal ions. Nevertheless, we must take account of the strengths (dissociation abilities) of functional groups and the fact that metal ions can be bound by other active centres, as studied in detail in [27]. It should be noted that IHSS published the content of acidic functional groups related to the content of carbon in humic acids and it is necessary to re-calculate the data on the whole samples of humic acids. On the other hand, this increase is not high, which means that only a smaller portion of metal present in hydrogel can be bonded by humic acids which corresponds with the low content of humic acids in hydrogel. No relationship exists between content of dyes in hydrogels and amounts of functional groups. There are more possibilities for the binding of dyes by humic acids. Apart from dissociable functional groups, the unsaturated and aromatic structures are more asserted due to the aromatic structure of studied dyes.

The knowledge of contents of diffusion probes was necessary for the mathematical description of their release from hydrogels. The effective diffusion coefficients *D*ef,h of Cu(II) ions and dyes in hydrogels were calculated on the basis of the following equation [26,36,37]:

$$m\_{\rm h\to s} = 2\frac{\varepsilon c\_{0,\rm h} - c\_{0,\rm s}}{1 + \varepsilon \sqrt{D\_{\rm s}/D\_{\rm cf,h}}} \sqrt{\frac{D\_{\rm s}t}{\pi}} \tag{1}$$

where *m*h→<sup>s</sup> is the total diffusion flux at time *t*; *c*0,h and *c*0,s are the initial concentrations of the probe in the hydrogel and aqueous solution (equal to zero in this case); *D*ef,h and *D*<sup>s</sup> are the effective diffusion coefficient of the probe in the hydrogel and the diffusion coefficient of the probe in the supernatant; ε is the ratio between concentrations of the probe in the supernatant (*c*s) and hydrogel (*c*h) in given time, i.e., ε = *c*s/*c*h.

The values of *D*<sup>s</sup> for Cu(II) ions are tabulated [38]: 1.43 <sup>×</sup> 10−<sup>9</sup> m2·s<sup>−</sup>1. The values of *D*<sup>s</sup> for dyes were determined in our previous study [22]. They were extrapolated for 25 ◦C and used in this work as: 8.42 <sup>×</sup> <sup>10</sup>−<sup>10</sup> <sup>m</sup>2·s<sup>−</sup>1for MB and 8.93 <sup>×</sup> <sup>10</sup>−<sup>10</sup> <sup>m</sup>2·s<sup>−</sup>1for RH [24]. These results are in agreement with values determined using other methods [16,19,39–41].

Experimental data fitted by Equation (1) are shown in Figure 1. We can see that they are in good agreement with the mathematical model. The slopes of the lines were used for the calculation of effective diffusion coefficients *D*ef,h. Their values are listed in Table 2.

**Figure 1.** Experimental data obtained for Cu (**black**), MB (**blue**) and RH (**red**) fitted by Equation (1).


**Table 2.** The values of effective diffusion coefficients of diffusion probes in the AG and AG–HA hydrogels.

The highest values of effective diffusion coefficients were determined for pure AG hydrogel. The obtained values can be compared with the results published in other studies [16,19,40,42–44]. Wang et al. [19] characterized the agarose hydrogel used in so-called DGT technique (diffusive gradients in thin films) for monitoring of different substances in natural environments (waters, soils, sediments). They investigated diffusivities of several ions including Cu(II) in 1.5 wt % agarose hydrogel and determined the value diffusion coefficient equal to 6.59 <sup>×</sup> 10−<sup>10</sup> m2·s−1. The value obtain in other study [43] for 2% agarose hydrogel was slightly lower (6.59 <sup>×</sup> 10−<sup>10</sup> m2·s<sup>−</sup>1). In this study, practically double value of *<sup>D</sup>*ef,h <sup>=</sup> 1.22 <sup>×</sup> <sup>10</sup>−<sup>9</sup> <sup>m</sup>2·s<sup>−</sup>1. This difference is partially caused by lower content agarose in our hydrogel and partially by different methods used for the determination of diffusivity, which is principal for the resulting value of the diffusivity [44]. The published values of diffusion coefficients of MB in 1.5 wt % agarose hydrogel (enriched by 3 wt % CaCl2) were between 2.9 and 3.9 <sup>×</sup> 10−<sup>10</sup> <sup>m</sup>2·s−<sup>1</sup> depending on the MB concentration and pH [40]. Similarly, the diffusion coefficients of RH in in 1.5 wt % agarose hydrogel were between 2 and 3.5 <sup>×</sup> 10−<sup>10</sup> m2·s−<sup>1</sup> depending on the concentration and pH [16]. Variations in obtained values of diffusion coefficients of RH were observed also for its diffusion in water (2.8–4.3 <sup>×</sup> <sup>10</sup>−<sup>10</sup> m2·s<sup>−</sup>1) [44].

In the case of release of Cu(II) ions, the highest value of *D*ef was determined for the AG–LEHA hydrogel. The same hydrogel achieved the highest initial content of Cu(II) ions. The LEHA sample can be characterized by the highest total acidity, and C/H and C/N ratios; aromaticity; and the lowest O/C ratio [34]. The lowest value of *D*ef was determined for the AG–ESHA hydrogel which can be characterized by the lowest C/N ratio, but C/H and O/C are comparable with LEHA. Simultaneously, ESHA has relatively high total acidity and aromaticity. The mobility of dyes in AG–ESHA and AG–LEHA hydrogels were similar. Both humic acids have high C/H and low O/C ratios. They are also more aromatic in comparison with other two samples. In contrast, the highest diffusivity of MB in the AG-PPHA hydrogel is probably caused by common impact of low content of acidic functional groups, C/H and C/N ratios and low aromaticity.

Chakraborty et al. [42] combined DGT with the CLE (competing ligand exchange) technique in order to investigate diffusion of metal ions in the presence of humic substances. They observed the increase in the diffusion coefficients from 6.06 <sup>×</sup> 10−<sup>10</sup> m2·s−<sup>1</sup> (obtained for Cu(II) ion) to 6.2 <sup>×</sup> 10−<sup>11</sup> m2·s−<sup>1</sup> (obtained for Cu–NLHA complex), 8.0 <sup>×</sup> 10−<sup>11</sup> m2·s−<sup>1</sup> (obtained for Cu–NLFA complex), and 8.5 <sup>×</sup> <sup>10</sup>−<sup>11</sup> m2·s−<sup>1</sup> (obtained for complex of Cu with Suwannee River natural organic matter). Similarly, the decrease in diffusion coefficient of Cu–HA in comparison with free Cu(II) ions in the hydrogel based on polyacrylamide cross-linked with an agarose derivative was from 5.48 <sup>×</sup> <sup>10</sup>−<sup>10</sup> to 5.70 <sup>×</sup> <sup>10</sup>−<sup>11</sup> m2·s−<sup>1</sup> [43]. In contrast, the effect of humic acids on the diffusion of RH in water was much weaker: from 2.88 <sup>×</sup> <sup>10</sup>−<sup>10</sup> to 2.22 <sup>×</sup> <sup>10</sup>−<sup>10</sup> m2·s−<sup>1</sup> for RH-PPHA complex and 2.15 <sup>×</sup> <sup>10</sup>−<sup>10</sup> m2·s−<sup>1</sup> for RH-SRHA complex. Their values of diffusion coefficient in 1.35 wt % AG hydrogels achieved 91%, 87% and 88% of diffusion coefficients in water, respectively [5]. It means that the effect of humic substances on the diffusion in agarose hydrogels observed by different authors differed. This finding showed that we must be very careful in the comparison of diffusion characteristics obtained by different authors and different methods [44].

The decrease of effective diffusion coefficients obtained for the hydrogels enriched by humic acids can be the result of two effects. The first is a possible change in hydrogel structure (see SEM of lyophilized hydrogels in Figure S3, from the Supplementary Materials). In spite of the fact that the content of humic acids in hydrogel is relatively low, their incorporation into AG hydrogel can influence its inner structure, including the distribution, size and shape of hydrogel pores [24]. The structure of humic acids is very dynamic and sensitive to circumstances such as concentration, pH and ionic strength [24,45,46]. They can be characterized by a supramolecular arrangement of relatively small particles in co-existence with bigger macromolecules [24,45–51], which makes it possible to respond to changes in their surroundings.

The second effect is a possible interaction between the diffusion probes and HA and humic acids during the transport of the probes through the hydrogel. In comparison with our previous works [21–28], the diffusion probes are in equilibrium with humic acids at the beginning of the release experiments. It means that the immobilization of the probe has the same rate as its liberation from binding sites. It is known that diffusion probes occurring in the hydrogels can be divided into three fractions: free mobile particles without chemical binding to humic acids, ion exchangeable ions bound by electrostatic forces, and strongly (covalently) bound particles in humic complexes [25,28,52]. These fractions are in a dynamic equilibrium and can convert to other forms as a result of changes in circumstances.

In the case of release experiments, the mobile fraction can easily diffuse out of the hydrogel. It results in the displacement from equilibrium, and particles of probe can be liberated from the exchangeable and strongly bound fractions. These processes can strongly affect the release of probe from AG–HA hydrogels and are dependent on the character of humic acids. In Figure 2, the ratios between effective diffusion coefficients *D*ef,h and the diffusion coefficients of probes in aqueous solutions *D*<sup>s</sup> and the ratios between effective diffusion coefficients *D*ef,h for AG–HA hydrogels and the values obtained for pure AG hydrogel are shown. The differences between dyes and Cu(II) ions were observed. While Cu(II) ions amount to 40%–80% of their diffusion coefficients in solution, that proportion is only 1%–6% in the case of dyes, mainly because of their sizes. The liberation of dyes from their ion-exchangeable and strongly bound fractions has an influence comparable with Cu(II) ions (AG–PPHA and AG–SRHA) or lower (AG–ESHA and AG–LEHA). The values of *D*ef,h obtained for hydrogels enriched by humic acids amount to 20%–70% of the values obtained for pure AG hydrogel.

**Figure 2.** The ratios between effective diffusion coefficients *D*ef,h and the diffusion coefficients of probes in aqueous solutions *D*<sup>s</sup> (**a**); the ratios between effective diffusion coefficients *D*ef,h for AG-HA hydrogels and the values obtained for pure AG hydrogel (**b**): Cu (black); MB (blue); and RH (red).

*Polymers* **2020**, *12*, 687

The influence of inner structure of hydrogel on the diffusion can be characterized as so-called structure fraction μ, which is the ratio between the porosity φ and tortuosity τ:

$$
\mu = \phi / \pi \tag{2}
$$

The value of μ can be determined as the ratio between the diffusion coefficient of the probe in AG hydrogel and the diffusion coefficient of the probe in water (*D*s). If we focus on the ratios obtained for pure AG hydrogel, we can state that their values are lower for the diffusion of dyes (5%–6%) in comparison with Cu(II) ions (> 80%). In the case of AG–HA hydrogels, the situation is more complex. The release of diffusion probes from hydrogels can be described as the non-stationary diffusion based on Fick's equation [36,37]:

$$\frac{\partial \mathcal{L}}{\partial t} = D\_{\text{ef},h} \frac{\partial^2 \mathcal{L}}{\partial \mathbf{x}^2} \tag{3}$$

where *c* represents the concentration of the diffusing compound at time *t* and position *x* (the coordinate parallel to the direction of the diffusion movement). The diffusion coefficient *D*ef,h is the main parameter characterizing the rate of the transport. The diffusion coefficient is an "effective" characteristic which reflects the influence of chemical interactions of diffusion probe with humic acids in their transport through the hydrogel and the influence of inner structure of hydrogel. Mathematically, the effects of the chemical reaction can be described by the following equation based on the conservation of mass:

$$\frac{\partial \mathcal{C}}{\partial t} = D^\star \frac{\partial^2 \mathcal{C}}{\partial \mathbf{x}^2} - \dot{r} \tag{4}$$

where *<sup>D</sup>*\* is the diffusion coefficient affected only by the porous structure of the hydrogel and . *r* is the rate of chemical reaction. In this case, the value of *D*\* is equal to *D*ef,h for pure AG hydrogel. If a fast chemical reaction in the presence of local equilibrium between free mobile probes (*c*) and immobilized ones (*c*im) is presumed (*K* is the equilibrium constant), then

$$
\sigma\_{\rm im} = \mathcal{K}c \tag{5}
$$

and Equation (4) can be written as

$$\frac{\partial \mathcal{C}}{\partial t} = D^\* \frac{\partial^2 \mathcal{C}}{\partial x^2} - K \frac{\partial \mathcal{C}}{\partial t'} \tag{6}$$

and consequently,

$$\frac{\partial \mathcal{L}}{\partial t} = \frac{D^\*}{1 + K} \frac{\partial^2 \mathcal{L}}{\partial \mathbf{x}^2} = D\_{\text{ef}, \mathbf{h}} \frac{\partial^2 \mathcal{L}}{\partial \mathbf{x}^2} \tag{7}$$

Since the diffusion coefficient in the hydrogel *D*\* is dependent on its porosity and tortuosity expressed by the structural factor μ according to the Equation (2), the following relation can be written:

$$D\_{\text{ef},h} = \frac{D^\*}{1+K} = \frac{\mu D\_{\text{s}}}{1+K} \tag{8}$$

in which the effects of the tortuous movement of the diffusing matter in the hydrogel and the chemical reaction between diffusion probe and humic acids are involved [25–28,36,37].

The values of *K* can be calculated only assuming that the inner structure of hydrogel was not changed by the addition of humic acids and they should be proportional to the ratios between effective diffusion coefficients *D*ef,h for AG–HA hydrogels and the values obtained for pure AG hydrogel shown in Figure 2b. As it was described in [24], rheological measurements showed that the AG hydrogel is more resistant to applied stress than hydrogels enriched with humic substances and the networks of the AG–HA hydrogels can easily collapse. The behaviour of hydrogels enriched with humic substances shifted towards that of viscoelastic liquids. This means that hydrogels containing humic substances had a lower ability to resist mechanical stresses, which can be connected with their higher permeability. Therefore the effect of interactions between diffusion probes and humic acids is probably higher than the values of *K* listed in Table 3.


**Table 3.** Values of the apparent equilibrium constants *K* determined on the basis of Equation (8).

The concentration profile in hydrogel during the release of Cu(II) ions and dyes can be described as [25,36,37]:

$$\mathbf{c} = \frac{1}{2} \mathbf{c}\_{0, \mathbf{h}} \left| err \frac{l - \mathbf{x}}{2 \sqrt{D\_{\text{ef}, \mathbf{h}} t}} + err \frac{l + \mathbf{x}}{2 \sqrt{D\_{\text{ef}, \mathbf{h}} t}} \right| \tag{9}$$

where *l* is the length of hydrogel and *x* is the distance from the interface between hydrogel and solution. This model is in a good agreement with data obtained for Cu(II) ions (see Figure 3). The small differences were observed close to the interface between hydrogels and solutions. The agreement between mathematical model and experimental data showed on the fact that the release of Cu(II) ions corresponded with our presumptions and they were accumulated on the interface. Similar results were obtained for all studied AG–HA hydrogels and Cu(II) ions. In contrast, the agreement of the Equation (8) with experimental data obtained for dyes was worse. It seems that dyes are accumulated in a certain distance from the interface. This was observed mainly in the case of RH (see Figure 3b). It is not easy to explain it. It is known that MB and RH can form bigger aggregates [53–57]. This formation together with general bigger size of dye probes (in comparison with metal ions) can support the observed accumulation.

**Figure 3.** The concentration profile of Cu(II) ions in AG–LEHA hydrogel after 40 h from the start of release (**a**), and the concentration profiles of MB (blue) and RH (red) in AG–LEHA hydrogel after 168 h from the start of release (**b**) fitted by Equation (8) – dashed curves.

As mentioned above, the transport through hydrogel can be affected by two factors: the tortuous movement of the diffusing particles in the porous structure of hydrogel and the interactions of diffusing particles with hydrogel. On the condition that pure agarose hydrogel cannot interact with a diffusion probe, we can determine the influence of the porous structure on the diffusion. The decrease in the diffusivity of probes in pure agarose hydrogels (in comparison with the diffusivity in water) can be attributed fully to the tortuosity effect. In general the movement of diffusing particles can be suppressed by their sizes. Particles of dyes are generally bigger than metal ions; therefore their Brownian motion is less intensive. On the other hand, pore size of agarose hydrogel exceed significantly the Stokes hydrodynamic radius of dyes [21,58]. The decrease is much stronger for dyes which can be connected with their sizes. The accumulation of dyes in a certain distance from the interface is more likely connected with the humic acids contained in enriched hydrogel. No accumulation was observed in the case of diffusion in pure agarose hydrogel. The most intensive accumulation was observed for the hydrogel enriched by LEHA, the weakest for ESHA. It is not easy to explain this finding. The phenomenon was observed only in release experiments. It means that dye was homogeneously distributed in hydrogel and (partially) complexed with humic acids and the equilibrium between humic acids, dye and formed complexes in the beginning is assumed. When the release of dye from hydrogel started, the equilibrium was distorted and some dye can be liberated from humic complexes. We assumed that the observed accumulation can be connected with the disruption of equilibrium and an effort of the system to attain new equilibrium. It seems that humic-dye complexes are (partially) able to diffuse towards the interface between hydrogel and water, but their movement is much slower in comparison with free dye particles. It resulted in the situation wherein an excess of free dye particles arose in the hydrogel closer to interface and their depletion in the hydrogel far from the interface; therefore, the equilibrium must be attained again and again as the release proceeds. This means that a part of free movable dyes can be complexed in the hydrogel closer to interface and a part of complexes can be disintegrated in the hydrogel far from the interface. Other effect is that the pores in hydrogel are filled by solution containing both free dyes and probably also by their complexes with humic acids which can obstruct the movement of smaller free dye particles. Both these effects probably resulted in the described state and the maximum observed on the concentration profile of dye in hydrogel. Different types of humic acids (extracted from different matrices) were used in order to compare their abilities to interact with diffusion probes and influence their release out of hydrogel. It is necessary to realize further experiments in order to investigate our findings in detail.

#### **4. Conclusions**

The influence of humic acids on the transport of metal ions and dyes in agarose hydrogel was studied. It was confirmed that humic acids retarded the transport of diffusion probes. Humic acids' enrichment caused decreases in the values of effective diffusion coefficients due to their complexation with diffusion probes. The effect of complexation was selective for the particular diffusion probe. The strongest effect was obtained for the diffusion of MB in the AG–SRHA hydrogel, the lowest one for the diffusion of Cu(II) ions in the AG–PPHA hydrogel. In general, the diffusion of dyes was more affected by the complexation with humic acids in comparison with metal ions. We assume that metal ions interacted preferentially with acidic functional groups. In contrast, dye can interact with acidic functional groups and the condensed cyclic structure of the dye probes supported their interactions with the hydrophobic domains of humic substances.

The results can be used in the investigation of the functioning of natural organic matter in the transport of pollutants in natural systems. Humic acids, as important constituents of soil organic matter, are able to affect, significantly, the migration and bioavailability of some pollutants in nature. In this study, agarose hydrogel was used as a model of a system with a homogeneous distribution of humic acids contaminated by metal ions and dyes. This model hydrogel was very wet in order to study the release of pollutants out of hydrogel. The purpose was to assess the effect of humic acids (as the constituent of soil organic matter) on the mobility of pollutants in wet soil. It means that pollutants present in soil can be partially complexed by humic substances and the movements of complexed and free pollutants are generally restricted if the soil is dry. In contrast, pollutants can diffuse relatively fast in wet soils, and their movement can be supported also by a convection in soggy soils. This study

was focused on the diffusion of pollutants in a model of wet soil (e.g., after rain). Results described in this study showed that some pollutants complexed by humic acids are (partially) able to diffuse through pore structure, but their movement is slower and can cause an accumulation of dyes in a certain position (distance from interface). This accumulation can influence the ensuing release from the pore structure into water (e.g., it can reduce the effective size of pores). Therefore, the results obtained in this study can help in the investigation of the functioning of natural organic matter in the transport of pollutants in natural systems. The effective diffusion coefficient determined on the basis of this study included the influence of pore structure and the interactions between humic substances and pollutants. Both pores and affinity of organic matter to pollutants differ with the type and quality of soil. Therefore, the methods and technics presented in this study can be used for predictions of the mobility and bioavailability of pollutants. The mathematical models for different diffusion processes are "universal" and can be used for different hydrogel materials (inert and reactive) and different diffusion probes. One of them could be the use of humic hydrogels as the material having controlled release of nutrients in agriculture.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/3/687/s1, Figure S1: Pure agarose hydrogel (left) and agarose hydrogel enriched by humic acids (right) in cuvettes, Figure S2: Schematic illustration of release experiment, Figure S3. SEM of lyophilized hydrogels (ZEISS EVO LS 10): pure agarose hydrogel (left) and agarose hydrogel enriched by humic acids (right).

**Author Contributions:** M.K. performed the experiments, processed and analysed the experimental data and wrote the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Programme for Sustainability I (Ministry of Education, Youth and Sports), grant number REG LO1211, Materials Research Centre at FCH BUT-Sustainability and Development.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Bovine Decellularized Amniotic Membrane: Extracellular Matrix as Sca**ff**old for Mammalian Skin**

**Andrea Catalina Villamil Ballesteros 1,\*, Hugo Ramiro Segura Puello 1, Jorge Andres Lopez-Garcia 2, Andres Bernal-Ballen 3, Diana Lorena Nieto Mosquera 1, Diana Milena Muñoz Forero 1, Juan Sebastián Segura Charry <sup>1</sup> and Yuli Alexandra Neira Bejarano <sup>1</sup>**


Received: 31 August 2019; Accepted: 23 November 2019; Published: 5 March 2020

**Abstract:** Decellularized membranes (DM) were obtained from bovine amniotic membranes (BAM) using four different decellularization protocols, based on physical, chemical, and mechanical treatment. The new material was used as a biological scaffold for in vitro skin cell culture. The DM were characterized using hematoxylin-eosin assay, scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR-ATR), and differential scanning calorimetry (DSC). The in vitro cytotoxicity of DM was evaluated using MTT. The efficacy of decellularization process was assessed through DNA quantification and electrophoresis. All the used protocols showed a high effectiveness in terms of elimination of native cells, confirmed by DNA extraction and quantification, electrophoresis, and SEM, although protocol IV removes the cellular contents and preserve the native extracellular matrix (ECM) architecture which it can be considered as the most effective in terms of decellularization. FTIR-ATR and DSC on the other hand, revealed the effects of decellularization on the biochemical composition of the matrices. There was no cytotoxicity and the biological matrices obtained were a source of collagen for recellularization. The matrices of protocols I, II, and III were degraded at day 21 of cell culture, forming a gel. The biocompatibility in vitro was demonstrated; hence these matrices may be deemed as potential scaffold for epithelial tissue regeneration.

**Keywords:** decellularization; biological scaffolding; bovine amniotic membrane; extracellular matrix; tissue regeneration

#### **1. Introduction**

Tissue engineering aims to regenerate damaged tissues, developing biological substitutes which along with a thriving cell growth, may restore, maintain, or improve a functional tissue [1–3]. This field has undergone rapid development in the last quarter of the twentieth century, although this science is devoted to skin regeneration it is still a major scientific and clinical challenge [4,5], and the healing response to chronic wounds is poorly understood and a matter of debate [6]. The skin can be considered as the largest organ, which covers the entire surface of the body and its main function is to serve as protective barrier against chemical, mechanical, and infectious damage. Nonetheless, injuries from trauma or skin-burns result in large-scale tissue loss, therefore, autografts, allografts, and xenografts are traditionally used. However, these kinds of treatments have limitations, such as immune rejection, and primary contraction [7].

Ideal skin substitutes should mimic the natural functions of the skin and the structural properties of the extracellular matrix; moreover, it has to protect the organism from protein loss, and it should improve the aesthetic appearance of the wound as well as inhibit the growth of exogenous microorganisms [7,8].

As an innovative treatment for skin injuries, biological substitutes have appeared, and they have the function of supporting growth, differentiation, and cell migration, and may come from different substrates either natural or synthetic, such as, collagen, gelatin, hyaluronic acid, fibronectin/fibrin, chitosan, alginate, polyglycolic acid, polylactic acid, and polycaprolactone; likewise, biological scaffolds composed of extracellular matrix (ECM) of decellularized tissues may be also used as biological substitutes [9–13]. Within this context, decellularized tissues and organs have successfully been used in a variety of tissue engineering/regenerative medicine applications, and the used decellularization methods vary as widely as the tissues and organs of interest [14]. The importance of ECM stems from its three-dimensional ultrastructure and its composition provides a microenvironment that guides the organization, growth, and differentiation of skin cells [15–21]. From all the mentioned substrates, collagen as a part of the ECM has been extensively employed as a biomaterial in cellular therapies and tissue engineering [14,21–24], and its relevance as a candidate for tissue engineering has been described in great extent [25,26].

Despite of the recent breakthroughs in terms of tissue regeneration, wound healing is a complex process that involves activation and synchronization of intracellular, intercellular, and extracellular mechanisms, including coagulation and inflammatory events, fibrous tissue accumulation, collagen deposition, epithelialization, contraction of the wound, tissue granulation, and remodeling [14,15,27,28]. Therefore, grafting materials must exhibit biodegradable, biocompatible, and adequate mechanical properties as well as support normal tissue regeneration [3,29]. In that regard, skin substitutes for wound healing from biological materials based on animal ECM have been developed and the decellularization process has reached an important level of success [14,15,30–33]. Over the years, xenografts have been obtained from various animal species, including birds, rodents, felines, canines, bovines, and swine [34,35]. To this day, the available acellular ECM scaffolds include swine and bovine equine substrates as well as human amniotic membranes [14,15,30,35–40].

Unfortunately, despite the numerous investigations in this area, clinical wound treatment remains unsatisfactory in many cases [15]; chronic wounds require long term and intensive care, and the associated cost are high [31]. Specialists agree that there is still no ideal skin substitute available [31,41,42] and the high costs and time required for the preparation of biological substitutes are crucial factors for developing new materials [35]. Moreover, the risk of zoonotic infections that might be transferred from the graft to the patient is latent, and allergic reactions is the main contraindication for using these kinds of materials [3,34,43]. On the other hand, the human amniotic membrane presents relevant disadvantages, such as that it is scarce due to its high cost, it has poor mechanical properties, and it may be a source of infectious diseases spreading [44,45].

The need for skin substitutes is of paramount importance specifically for large defects of burns, congenital diseases, traumas, and infections [46]. In this frame, acellular amniotic membrane might have potential as a matrix for tissue regeneration or as a substrate to facilitate autologous/allogeneic cell transfer [47]. Human decellularized amniotic membrane has been widely shown as a biodegradable and bioactive matrix for regenerative tissue repair [48]. Thus, this study proposes a candidate that brings the inherent attributes of bovine amniotic membranes (BAM), which can be useful for being used as a matrix for skin regeneration. For this purpose, four distinct protocols were proved (details are shown in methodology section). The obtained decellularized membranes (DM) were used as an alternative scaffold for skin regeneration, and as a sources of collagen IV and VII, elastin, laminin one and five, fibronectin, and entactin [46,49,50]. This material has similar properties to other matrixes obtained from skin [51,52]. Therefore, DM were obtained from four different methods and the efficiency of these methods were evaluated. The cell cultures were carried out on samples and there is evidence of the material does not behave cytotoxically. The obtained DM demonstrated potential in the use of skin regeneration, which is

a valuable alternative for tissue engineering and the prospectives of its applications are a new challenge in the field of biomaterial science.

#### **2. Materials and Methods**

#### *2.1. Decellularization of the BAM*

The bovine amniotic membrane was obtained from a vaginal birth of a bovine female with no infectious-contagious diseases within aseptic conditions. Samples were collected and transported at 4 ◦C in centrifuge tubes containing a solution of phosphate buffered saline (PBS) that included antibiotics (penicillin, streptomycin, and amphotericin B). Thereupon, the samples were washed with cold PBS and dissected in sections of 16 cm2 sections in the biological class II biosafety cabinet (ESCO, IDN).

Four protocols were used to decellularize the BAM (Table 1). A control sample was kept without frozen treatment at −20 ◦C. For protocols I, II, and III, the separation of two layers, fetal and maternal, was performed as it is reported in most of the investigations [52–55]. On the other hand, this procedure was not performed in protocol IV, in order to observe variations in the properties of the membrane regarding exposure with the chemical solutions used and in cell culture.


**Table 1.** Decellularization protocols for the bovine amniotic membranes (BAM).

All BAM were subjected to a freezing cycle in liquid nitrogen (−196 ◦C) for 22 h and unfreezing in a serological bath (Polyscience, Niles, Illinois, USA) at 37 ◦C for two hours. Then, BAM were treated with strong and weak detergents (sodium dodecyl sulphate (SDS) 0.1% or Tween 80) for 4 h followed by being soaked in a base solution (NaOH 0.1 M) for 1 h and acid solution (peracetic acid (PAA) and ascorbic acid or ethanol). After, as a final wash, ethanol at 70% was applied for 1 h to remove residual nucleic acids and phospholipids from the tissue and finally, PBS as a buffer solution was pertained for 2 h. The membranes were mechanically stirred throughout the process using an orbital shaker (Camlab, Cambridge, UK) to ensure a homogeneous wash and a minimal damage to the tissue ultrastructure [5–7,30].

Once the abovementioned process was finished, protocols II and IV needed a new acid/basic treatment in order to wipe out the remains of color in the membranes; therefore, it was necessary to immerse them again in NaOH for one hour and PAA for another hour. For BAM treated in protocols I and III, it was not necessary to carry out more washings, and samples were stirred in ethanol at 70%.

After each decellularization step, BAM were washed with deionized water for 30 min in a shaker to eliminate tissue remnants and the used substances. Finally, all the membranes were washed four times with PBS for 30 min and stored at −22 ◦C.

#### *2.2. Determination of DNA Content*

#### 2.2.1. Extraction of DNA

To ensure the removal of all cellular and nuclear material in the decellularized BAM, the DNA extraction process was carried out using a PureLink® kit (Invitrogen). <25 mg of BAM and DM were placed into a micro centrifuge tube. It was added to 180 μL of genomic digestion buffer and 20 μL of proteinase K to remove lipids and digest proteins. Then, the treated samples were incubated at 55 ◦C in a serological bath with vortex every 10 min for one hour and centrifuged at 13,000 rpm for 3 min at room temperature. Each supernatant was transferred to a new sterile micro centrifuge tube and 20 μL of RNase A was added, mixed, and incubated for two minutes. Subsequently, 200 μL of lysis buffer and 200 μL of 99.9% ethanol were added to each lysate to precipitate the DNA by vortexing for five seconds.

Once the DNA was extracted, it was purified by a series of washes, placing the previous preparations in collector tubes with a column and centrifuging at 12,000 rpm for one minute. Washes were carried out with 500 μL of wash buffer, 1500 μL of wash buffer two, and 50 μL of buffer elution, centrifuging after each addition for one, two, and three minutes, respectively. Finally, the microcentrifuge tubes containing DNA were stored at 4 ◦C.

In the wells of an agarose gel, samples of the extracted DNA were placed in the horizontal electrophoresis chamber (Thermo EC, Holbrook, NY, USA). Afterwards, the movement of the bands was observed in the transilluminator (Fisher Biotech, Pittsburgh, PA, USA).

#### 2.2.2. DNA Quantification

DNA concentrations were obtained using a QuantiFluor® dsDNA System Kit (Promega, Madison, Wisconsin, USA). After the DNA extraction, the DNA samples were prepared by the addition of 1–20 μL to 200 μL of working solution in 0.5 mL PCR tubes and vortexing, and incubated at room temperature for five minutes, in a dark condition. Finally, fluorescence was measured in the calibrated Quantus™ fluorometer (Promega, Madison, Wisconsin, USA). The effectiveness of each decellularization protocol was evaluated by triplicated.

#### *2.3. Cell Culture*

Cells were obtained from a full thickness ovine skin biopsy and cultured with RPMI-1640 culture medium, supplemented with 5% fetal bovine serum (SFB) and 1% antibiotic penicillin, streptomycin, and amphotericin B (Sigma-Aldrich, Bogota, Colombia) in an incubator at 37 ◦C in a humidified atmosphere of 5% CO2. Monitoring culture was carried out every three days with an inverted microscope (Olympus, New York, NY, USA). After cell confluence, cells were sub-cultured through trypsinization and used in the second pass.

Cell Seeding on DM and 3-(4,5-d imethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) Assay

The in vitro cytotoxicity of DM was evaluated using 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazoliumbromide (MTT) assay (Cell Biolabs Inc., Bogota, Colombia). Specimens of 10 mm in diameter were cut and placed at the bottom of a 24-well-plate (Corning, New York, NY, USA). Before the

culture, samples underwent a sterilization process with ultraviolet light for 15 min each side. Then, cells were seeded on DM at a density 5 <sup>×</sup> 104 cells/well in the previous mentioned incubation conditions.

After culturing for 24, 48, and 72 h, 50 μL of the CytoSelect™ MTT Cell Proliferation Assay Reagent was added to each well and incubated for 4 h, until purple precipitate was visible. Then, 500 μL detergent solution was added and incubated at room temperature for two hours. A specific culture media (RPMI) was also considered as control. The absorbance of solution was measured using a microplate reader 800 TS (Biotek, Winousky, Vermont, USA) at 490 nm. Cell viability was determined using Equation (1). For histological analysis with hematoxylin-eosin, a DM sample was cultured until day 21.

$$\text{Cell viability (\%)}=\text{Absorbance sample} / \text{Absorbance control (untreated)} \times 100\tag{1}$$

#### *2.4. Histological Analysis with Hematoxylin-Eosin*

Control sample BAM, DM, and recellularized DM (at 21 days culture) tissues were fixed in 10% formaldehyde, embedded in paraffin, cut into sections of 5 μm, stained with hematoxylin-eosin, and observed under the optical microscope (Olympus, Tokyo, Japan) to evaluate the presence of nuclear material.

#### *2.5. Scanning Electron Microscopy (SEM)*

Micrographs of the prepared samples were taken by the scanning electron microscope Nova NanoSEM 450 (FEI, Brno, Czech Republic) with a Schottky field emission electron source operated at an acceleration voltage ranging from 200 V to 30 kV and a low-vacuum SED (LVD) detector. A coating with a thin layer of gold was performed by a sputter coater SC 7640 (Quorum Technologies, Newhaven, East Sussex, UK).

#### *2.6. FTIR-ATR Spectroscopy*

FITR spectroscopy analysis was carried out on NICOLET 6700 FTIR spectrometer device (Thermo Scientific, Waltham, MA, USA) equipped with attenuated total reflectance (ATR) accessory utilizing the Zn–Se crystal and software package OMNIC over the range of wavelengths from 4000 to 600 cm−<sup>1</sup> at room temperature under a resolution of 4 cm<sup>−</sup>1. Each spectrum represents 64 co-added scans referenced against an empty ATR cell spectrum.

#### *2.7. Di*ff*erential Scanning Calorimetry (DSC)*

Calorimetric measurements were carried out in a differential scanning calorimetry (DSC) 1 calorimeter, Mettler Toledo (Greifensee, Zurich, Switzerland), under nitrogen flowing at a rate of 30 mL min−1. The specimens were pressed in sealed aluminum pans. A heating cycle was performed in order to acquire the glass transition temperature (T*g*) and melting temperature (T*m*). The samples were cooled down by nitrogen at an exponentially decreasing rate. The heating of the cycle was performed from 25 to 240 ◦C at a rate of 20 ◦C/min. The T*<sup>g</sup>* was determined as the midpoint temperature by standard extrapolation of the linear part of DSC curves using Mettler-Toledo Stare software and the T*<sup>m</sup>* as the maximum value of the melting peak.

#### *2.8. Statistic Analysis*

MTT measurements were performed in triplicate. All experimental values were expressed in form of average ± standard deviation. Results were statistically compared using one-way analysis of variance (ANOVA) with *p* < 0.05.

#### **3. Results and Discussion**

#### *3.1. Decellularization of BAM and DNA Content*

The main purpose of decellularization of xenogeneic matrices is to effectively eliminate cells and nucleic acid residues, as well as preserve the composition of the ECM [11]. In this frame, the DNA content analysis in DM indicated total cell absence whereas in BAM is easily observable (Figure 1). Moreover, the four protocols achieve cell removal until the detection limit of the test.

**Figure 1.** Electrophoresis pattern obtained in agarose gel for the used protocols (I, II, III, and IV), and for the control sample (BAM).

In the electrophoresis technique, the DNA moieties are so small that they cannot be observed while they migrate through the gel, as it is possible in the control membrane (BAM). This technique allowed the separation, identification and isolation of DNA fragments, which cannot be separated by other methods. However, the quantification of DNA allows for the detection of small amounts of the nucleic acid, that is, the actual value of respective DNA for each protocol. From smallest to largest value, so it is highly sensitive. DNA content analysis of DM was conducted to compare the efficiency of the previously developed decellularization protocols. The obtained results showed that the DNA levels decreased with each protocol in comparison to BAM (Figure 2). ANOVA test showed that no significant differences were evidenced in the tested protocols. It has been reported that a lower concentration of <50 ng/mg in a membrane implies that the matrix can be considered as decellularized [32,56]. Therefore, the chemical and mechanical methods used were effective at eliminating the DNA content from the DM. Other studies have shown that higher degrees of decellularization measured by DNA content are associated to a better tissue remodeling in vivo and macroscopic response in the host [57,58]; therefore, protocol II would be considered as the most suitable for decellularization process.

The decellularization protocols consisted of the application of physical freeze-unfreeze method to lyse cells through the formation of microcrystals. This technique requires smaller amounts of chemical agents, which do not significantly alter the ECM properties [10,59,60]. With liquid nitrogen, a lower number of cycles and shorter time were required compared to freezing-unfreeze protocol at −20 or −80 ◦C [60].

Sodium dodecyl sulphate used as an ionic detergent has the ability to efficiently remove cells and genetic material [58,61] as it was observed in the electrophoresis and confirmed by the DNA content of protocols I and II. Likewise, SDS contributes to the inhibition of collagen calcification processes [62]. However, it can alter the ultrastructure and the elimination of growth factors [58,59,61]. Tween-80 is a non-ionic detergent, considered mild, that has the property of solubilizing proteins while maintaining

the structure of the native protein [10]. In the electrophoresis of samples prepared using protocols III and IV, a slight sweep was observed due to protein residues most likely associated with the use of this detergent and DNA content analysis corroborated the presence of DNA in low concentration after decellularization. DM from protocol IV contained double layer (amnion and chorion), therefore, the surface area of exposure to chemical agents was smaller and consisted of even more DNA residues. For this reason, it contains more DNA; however, the obtained value for this protocol is lower in comparison to the reported value of <50 ng/mg for a membrane which is considered decellularized.

**Figure 2.** DNA content obtained for the control sample and decellularized membranes (DM).

It is imperative to emphasize that the use of alkaline or acidic solutions in excess may cause serious alteration on the ECM [10]. Alkaline solutions denature chromosomal DNA and plasmid; however, they degrade collagen to a certain extent and eliminate growth factors from the resulting DM and reducing its mechanical properties. The exposure of the matrices to NaOH was performed for one hour, since a prolonged exposure may disintegrate the tissues and interrupt the formation of collagen crosslinks [3]. On the other hand, acids dissociate the DNA of the ECMs via solubilization of cytoplasmic components and the disruption of nucleic acids. PAA with hydrogen peroxide or ethanol was effective for the disinfection and removal of cellular debris from the BAM [3]; ethanol was used for the final wash to eliminate the residual nucleic acids, and delipidize the tissues in addition to its microbicide action, necessary for the manipulation to which the membranes were exposed [10]. Washes with PBS were indeed effective to remove chemical traces and to neutralize the pH of the samples for cell culture. Cells need strict culture conditions to survive, and variation in those conditions can trigger apoptosis [63].

In comparison to other human amniotic membrane decellularization studies, no antibiotics or enzymes were used, which are usually associated with bacterial resistance and irreparable damage to the matrices [53]. Finally, the low obtained standard deviation in this process is an indicative of the reproducibility of the decellularization protocols.

#### *3.2. MTT Assay*

The viability of skin cells seeded on DM of different protocols was measured in terms of cellular mitochondrial dehydrogenase activity using MTT assay. The viability of seeded cells for 24, 48, and 72 h are depicted in Figure 3. It was observed that the cells sustained their metabolic activity in culture on the DM, and that activity was increased during the time, showing a considerable biocompatibility of DM. At 72 h, the activity was not observable as a consequence of the detergent did not longer dissolve MTT.

**Figure 3.** 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) assay of skin cells growing on different DM for 24, 48, and 72 h. \* *p* < 0.05.

The obtained data underwent ANOVA testing and the results indicated that no significant differences were evidenced in all the tested protocols. This was likely due to the nature of DM and and its composition (mainly collagen). Moreover, it was revealed that detergent, acids, and bases removal are critical for generating optimal acellular scaffolds with potencial clínical uses. In this way, any of the tested protocols show cytotoxic effects on the seeded cells. The cell number augmented with increasing the incubation time, which is an indicator for improving the effect of DM on the metabolic activity of the cells compared to a control culture. DM of protocol IV exhibited a visible increase in metabolic activity, associated to the fact that it retained its biochemical properties to a greater extent. These results are in a good agreement with other reports which indicate that scaffolds made of decellularized amniotic membrane, did not exhibit cytotoxicity [64].

The results of the previous studies suggest that the vast majority of current decellularization protocols are detergent-based and incompletely removed residual detergents may have a deleterious impact on subsequent scaffold recellularization [10,29,58,65]. Residual SDS within biomaterials has severe cellular toxicity and may be responsible for the decrease in cell growth [10,29]. Therefore, the success of subsequent recellularization is based on the removal of the lysed cellular material and cytotoxic detergent after the decellularization process [65].

The progress of cell cultures is shown in Figure 4. On day 21 in protocols I, II, and III, degradation was observed and the membrane of protocol IV remained intact. Furthermore, protocols I, II, and III were degraded and it was not possible to carry out histological analysis. The histological findings corroborated the cell growth on the DM.

**Figure 4.** Images of cell culture on DM for mammalian skin cells obtained using an inverted microscope (40×).

#### *3.3. Histological Analysis*

In the histological study of BAM, a simple cubic epithelium was observed (Figure 5) that included large binucleated (basophilic) cells and native collagenous (eosinophilic) fibers [49,54] of normal bovine tissue. In this technique, the efficiency of the decellularization protocols was substantiated by cellular absence. Hematoxylin-eosin assay (H&E) disclosed abundant mammalian skin cells adhered to the recellularized BAM as it was observed during cell culture monitoring until day 21. The microphotographs are shown in Figure 5. No cells were observed in the DM as a consequence of the acidophilic matrix. Moreover, cells were present after 21 days of culture in the recellularized DM.

**Figure 5.** Microphotographs for BAM (**left**), DM (**middle**), and recellularized DM (**right**) obtained from the hematoxylin-eosin (H&E) assay (100×).

Although several studies recommend the use of cell lines in this kind of experiments [66–71], it is of paramount importance to indicate that in vitro studies have evidenced that in a standard cell culture, fibroblast positively influence keratinocyte growth, most likely due to the fact that these cells secrete soluble growth factors. In natural skin, the interaction is relevant as well. Without fibroblasts, the keratinocyte differentiation is severely affected. Moreover, keratinocytes have also a positive effect on the proliferation of fibroblasts. Based on these findings, it is possible to affirm that in order to gain meaningful data from toxicological in vitro studies, the isolated focus on a keratinocyte-containing epidermal layer alone is not sufficient, making the use of a full-thickness skin model essential [14,72–74].

The amniotic membrane has structures, which are histologically similar to the skin, i.e., composed of a multilayer epithelium and the basic membrane, and the structure might be considered as a good support for wound healing, reepithelialization and inhibition of scar formation and bacterial growth [50,52,75,76].

#### *3.4. Scanning Electron Microscopy*

Topographical analysis shows that the native BAM contained collagen fibers with tissue cells on an irregular surface (Figure 6). This result is in a good agreement with the obtained by electrophoresis and histology studies, where the control membranes presented the DNA band.

**Figure 6.** SEM image for BAM.

The micrographs of the studied membranes also confirm that the processes were efficacious for the elimination of the cells in all the tested protocols. There are differences in the surface of each membrane; for instance, image from protocol I depicts a surface where the collagen fibers are very similar to the native ones, whereas DM for protocol II is a smoother surface. DM obtained using protocol III showed tissue wear along with some crystalline residues and the sample of protocol IV is the most homogeneous of the appraised surfaces (Figure 7).

**Figure 7.** SEM images for DM obtained from the tested protocols.

#### *3.5. FTIR-ATR Spectroscopy*

Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra of the assessed samples are shown in Figure 8. The peptide characteristic bands at approximately 3300, 3000, 1630, 1545, 1240, and 690 cm−<sup>1</sup> are identifiable. For example, the amide I is a broad band around 1640–1630 cm−<sup>1</sup> originated from C=O stretching vibrations coupled to N–H bending vibration. The amide II band, which is located at around 1550 cm−<sup>1</sup> arises from N–H bending vibrations coupled to C–N stretching vibrations. Finally, the amide III characteristics bands, that usually appear within the range of 1300–1200 cm−<sup>1</sup> result from the interaction between N–H bending and C–N stretching. The band locate at 690 cm−<sup>1</sup> is an usual amide vibration which emerges from out of plane N–H wagging [77–80].

**Figure 8.** FTIR spectra for DM obtained from the four protocols.

The absorption peaks within the 3000–2800 cm−<sup>1</sup> spectral range are attributed to aliphatic C–H stretching; on the other hand, the bands around 1500 cm−<sup>1</sup> are associated with C–H bending. The studied spectra possess the typical features of collagen-like proteins, which have been extensively studied in previous scientific works [77,78,81]. Nevertheless, the characteristic collagen bands are visible, which implies that collagen is retained upon each decellularization process, there are visible differences in the intensity of spectral bands, which may be ascribed to the interaction of the membranes with the solutions, the duration, and harshness of each decellularization protocol. In fact, there is a triple helix denaturation, which may be evinced on the intensity bands.

#### *3.6. Di*ff*erential Scanning Calorimetry (DSC)*

The thermograms of the decellularized membranes of protocols II, III, and IV are shown in Figure 9, and they correspond to the first heating scan. It was not possible to obtain the thermogram from the DM of protocol I because of the sample decomposition. Different endothermic denaturation peaks may be seen viz.: II, 185 ◦C; III, 200 ◦C; and IV, 152 ◦C. Collagen materials exposed to high temperatures endure irreversible denaturation process [81–86]. Previous thermo-analytical studies of denaturation of collagen report that the denaturation temperature for bovine skin is 55 ◦C [87], bovine intramuscular connective tissue 90 ◦C [88], rat tail collagen 65 ◦C [89], bovine skin 50–55 ◦C [84], and type I collagen from bovine skin soluble in acid 117 ◦C [90]. As the water content is higher, collagen denaturation temperature gets higher, and this phenomenon may be observed in this study. It should be noted that collagen denatures; therefore, there was no second heating scan; furthermore the treated samples also showed another endothermic peak (105–115 ◦C) which is most likely related to gelatinous structures by the denaturation that were obtained at day 21 of culture for protocols II and III [84].

**Figure 9.** Thermograms for DM obtained from four different protocols.

#### **4. Conclusions**

Decellularized membranes have attracted the attention of the scientific community since through tissue engineering it is possible to develop biological scaffolds that aim to deliver cells and proteins to damaged tissue, and at the same time, gradually degrade to make room for regenerated tissue. Within this frame, BAM were decellularized using four different protocols and the differences in terms of decellularization can be considered as negligible. All membranes obtained DNA concentrations <50 ng/mg, indicating that traces of the nucleic acid were present in the prepared material, although the obtained values are negligible which implies that DM do not have presence of native cells from the BAM. Nonetheless, protocol II proved to be the best method in terms of eliminating DNA content.

In the biological test, the obtained matrices from BAM were not cytotoxic for the cells (confirmed by MTT) and consisted of a source of collagen for recellularization. The mammalian skin cells adhered and conducted a remodeling effect on the BAM.

Each protocol may damage the ultrastructure of the tissue in different grade, mainly related to the chemical substances that were used. The extent of denaturation depends upon the interaction of the chemical substances with the molecules present in the tissue, and this analysis was supported by spectroscopic, thermal and topographical techniques.

Results showed that protocol IV (SDS 0.1%, NaOH 0.1 M, PAA + ascorbic acid 0.1, ethanol 70%, and PBS) could efficiently remove the cellular contents and preserve the native ECM architecture (confirmed for FTIR-ATR spectroscopy). Therefore, double layer bovine amniotic membranes (fetal and maternal) retained its biochemical properties after decellularization in comparison with the other membranes. Moreover, the mentioned double layer membrane exhibits a very low DNA concentration which is below to 50 ng/mg; for this reason, DM of protocol IV might be used as a possible biological substitute for skin.

The membranes of protocols I, II, and III, being single layer (stromal), had a greater surface area of exposure to the chemical agents used and, therefore, degraded further in terms of their composition. However, degradation was observed in culture, a semi-transparent gel was formed that may have potential biomedical applications, which may be part of later studies, underlining the potential applications of this matrices for tissue engineering.

The in vitro biocompatibility was demonstrated in this study, and it is indeed of pivotal importance, since this matrix may be considered as a potential source for the regeneration of epithelial tissue.

**Author Contributions:** A.C.V.B. and H.R.S.P. designed the study and carried out the biological experiments. J.A.L.-G. and A.B.-B. were responsible for the characterization techniques. D.L.N.M., D.M.M.F., and Y.A.N.B. collaborated in the biological test. A.C.V.B., J.A.L.-G., and A.B.-B. wrote and edited the manuscript. J.S.S.C. collaborated by reviewing the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** We would like to thank to Juan Villamil Fraile for donating the BAM; to Andrés Gutierrez Beltrán, Eng. Sergio Andrés Pineda, and Clínica Veterinaria Potencia Animal for their support in data analysis. Special thanks go to the program *Reconocimiento Docente* for facilitating a posdoctoral visit to Tomas Bata University

in Zlin and to Universidad Distrital Francisco José de Caldas, where part of the characterization techniques was carried out.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Atmospheric Pressure Microwave Plasma Jet for Organic Thin Film Deposition**

**Mehrnoush Narimisa 1,\*, František Krˇcma 2, Yuliia Onyshchenko 1, Zdenka Kozáková 2, Rino Morent <sup>1</sup> and Nathalie De Geyter <sup>1</sup>**


Received: 29 November 2019; Accepted: 28 January 2020; Published: 6 February 2020

**Abstract:** In this work, the potential of a microwave (MW)-induced atmospheric pressure plasma jet (APPJ) in film deposition of styrene and methyl methacrylate (MMA) precursors is investigated. Plasma properties during the deposition and resultant coating characteristics are studied. Optical emission spectroscopy (OES) results indicate a higher degree of monomer dissociation in the APPJ with increasing power and a carrier gas flow rate of up to 250 standard cubic centimeters per minute (sccm). Computational fluid dynamic (CFD) simulations demonstrate non-uniform monomer distribution near the substrate and the dependency of the deposition area on the monomer-containing gas flow rate. A non-homogeneous surface morphology and topography of the deposited coatings is also observed using atomic force microscopy (AFM) and SEM. Coating chemical analysis and wettability are studied by XPS and water contact angle (WCA), respectively. A lower monomer flow rate was found to result in a higher C–O/C–C ratio and a higher wettability of the deposited coatings.

**Keywords:** atmospheric pressure plasma jet (APPJ); microwave (MW) discharge; thin film deposition; optical emission spectroscopy (OES); Comsol MultiPhysics; methyl methacrylate (MMA); styrene

#### **1. Introduction**

Atmospheric pressure non-thermal plasma deposition as an advantageous coating technique has been a continually growing research field for over a few decades [1–3]. This method has already been frequently used to deposit a wide range of convenient thin films that were previously fabricated by various other methods such as chemical synthesis, electrochemical polymerization, and low pressure plasma-enhanced chemical vapor deposition (PECVD) [4–6]. When using the latter methods, controlling the deposition rate, optimizing the structure and adhesion of the produced polymers, reducing the energy consumption, and the enormous cost of vacuum techniques were remaining concerns. Consequently, the field of thin film deposition remains open to alternative methods that can overcome these issues such as atmospheric pressure plasma deposition. Among atmospheric pressure plasma sources suitable for coating deposition, atmospheric pressure plasma jets (APPJs) have gained recognition due to their straightforward design and operation together with their ability to effectively coat 3D complex substrates and their capacity to produce a localized deposition [7–11]. Furthermore, due to its construction, a plasma jet offers the possibility of transporting large amounts of charged and active species while the sample is in indirect exposure to the active discharge, which in turn is helpful for biomedical purposes as well as for the treatment of other sensitive materials [12–15]. The concentration of reactive species in the plasma directly correlates with the discharge excitation

method. The majority of studied atmospheric pressure plasma jets are driven by audio- (typically around 10 kHz) or radiofrequency (usually 13.54 or 27.2 MHz) power supplies and contain a large amount of reactive species [16–19]. Discharges, generated with even higher excitation frequency (e.g., microwave (MW)), contain an even higher concentration of these reactive species [20,21]. In the case of an MW discharge, the plasma can be (as in the present case) created using the surface wave that propagates along the plasma–capillary or plasma–open-air boundary, and, thus, the active discharge propagates out of the capillary and can also interact directly with the surface. Due to this fact, in an MW-induced discharge, the radicals are present at a high density, thereby contributing to a raised level of fragmentation of the introduced monomer while maintaining room temperature. The plasma torch length can be varied by gas flow and supplied energy [22] and the monomer can be mixed with plasma at different positions with respect to the end of an active discharge. The high radical density in an MW-induced discharge is very advantageous as a high rate of chemical dissociation accompanied by subsequent recombination processes is a crucial reaction for thin film deposition. Furthermore, such a relatively high frequency can also minimize the plasma instability and its transition to arc mode during film deposition, which can bypass the degradation of molecules and chemical bonds during the deposition stage. Nevertheless, despite all the above-listed advantages of MW discharge sources, only a few works have been conducted which have focused on the performance of MW plasma jets in thin film deposition under different operational conditions [8,23,24]. Additional advantages of MW plasma application for thin film deposition are its simplicity and its generally cheap scalability up to a couple of meters. Thus, laboratory achievements could be generally simply extended to industrial scale practice.

A thorough investigation of a complete experiment is required to give a full insight into the processes involved in plasma jet polymerization and properties of the deposited films. Plasma diagnostics in this work are used as a powerful tool to establish the connection between monomer fragmentation/distribution and coating properties, as it is crucial to demonstrate correlations between the plasma discharge and the resultant coatings. This information is profoundly beneficial for future plasma source development and for tuning the experimental conditions in order to obtain plasma-polymerized coatings with specific requirements suitable for a particular application.

Many different monomers can be used for organic thin film deposition by an MW plasma jet. However, each particular monomer has a specific chemical structure, and, thus, requires a particular energy of the plasma reactive species to be successfully deposited as a coating on a substrate. Among the available monomers, methyl methacrylate (MMA) and styrene are very interesting from a technological point of view. A wide range of applications such as humidity sensors, optical biosensors, protective coatings, biomedical applications, and anti-reflection optical coatings require MMA and/or styrene-based coatings because of their ideal plasticity, transparency, electric insulation, and cushioning properties [25–28]. Taking this into account, this work focuses on the optimization of styrene and MMA coating deposition by an APPJ operating at an MW frequency of 2.45 GHz. This plasma source has been recently developed and has never been applied for thin film deposition before [20,29]. Consequently, this work is a unique example of using an MW source to polymerize styrene and MMA. Initially, the reactive species in the plasma source are characterized using optical emission spectroscopy (OES). In the next step, gas flow dynamic simulations with Comsol MultiPhysics 5.3 are performed to demonstrate the distribution of the monomer during the deposition process [30–32]. In doing so, the gas mixing and propagation in the working volume can be prognosticated, which can in turn help to understand the processes that are taking place during plasma deposition. Moreover, the results of the gas dynamic simulations extensively help to interpret other experimentally obtained data such as temperature and incorporation of other species, which can significantly influence the final coating properties [33,34]. In the final part of the study, the physical and chemical properties of the deposited films are also characterized by water contact angle (WCA) analysis and XPS, while the surface morphology and topography are analyzed using SEM and atomic force microscopy (AFM), respectively. By following

this research strategy, crucial information regarding atmospheric pressure MW plasma jet effectiveness for thin film deposition will be revealed.

#### **2. Materials and Methods**

#### *2.1. Materials*

Plasma deposition experiments were carried out using styrene (99.5%, Brenntag CZ) and MMA (99.9%, Pliska, Brno, Czech Republic) as precursors. Cover slip glasses 25 mm in diameter were selected as substrates. Furthermore, to have a better visualization of the coating lateral dimensions, silicon wafers (Siegert Wafers, Aachen, Germany) were also used as substrates. Argon of 99.996% purity purchased from Linde was used as the main and carrier gas in the discharge.

#### *2.2. Experimental Plasma Set-Up*

A schematic representation of the experimental plasma set-up is depicted in Figure 1. Similarly to a previous study [29], a high voltage power source (Sairem, GMS 200 W) operating with low power at a 2.45 GHz frequency was used for plasma generation. This power supply was connected to a surfatron surface wave (SW) launcher (Sairem, SURFATRON 80, Décines-Charpieu, France) using a coaxial cable. A continuous flow of water was utilized to cool down the surfatron resonator to avoid the influence of elevated temperature. Plasma was generated in a flow of 1 standard liter per minute (slm) argon gas using an Omega FMA-A2408 mass flow controller (Omega, Norwalk, CT, USA) inside a quartz capillary with a 3.06 and 8.0 mm inner and outer diameter, respectively. To introduce the monomer into the plasma, a Macor disc (60 mm diameter and 6 mm thickness) with a cylindrical opening 2 mm in diameter at one side for the monomer carrier gas inlet was placed at the end of the capillary edge. In the middle of this disc there was also a circular opening 4 mm in diameter which ensured that the plasma effluent could freely pass through this opening. Due to the geometry of the MW plasma jet set-up, the precursor was inserted in the plasma afterglow region perpendicular to the main gas stream and towards its core. The Macor disc was also able to improve the efficiency of deposition due to the fact that it alters the gas dynamics in the volume close to the substrate, as previously reported by Onyshchenko et al. [35]. To deliver different amounts of evaporated monomer molecules into the plasma plume, a Bronkhorst F201 mass flow controller was used to regulate an additional argon gas flow, as a monomer carrier gas, with different flow rates (125, 250, 500, and 750 standard cubic centimeters per minute (sccm)) through a glass bubbler filled with MMA or styrene. The bubbler was kept in a water–ice mixture to keep the temperature of the liquid monomers constant. During the plasma deposition experiments, the distance between the bottom edge of the disc and the sample was varied between 3, 5, and 7 mm. For these three distances, the active discharge part of the plasma effluent did not touch the substrate. The other plasma operational parameters that were varied in this work were applied power and deposition time; these were 7, 10, and 13 W and 1, 3, and 5 min, respectively. This work aimed to demonstrate the influence of each of these working parameters on plasma jet performance in terms of thin film deposition. For sake of clarity, one set of experimental parameters ("Main") was selected as a reference point for all other parameters that were varied. The 10 sets of plasma operational parameters used in this study are summarized in Table 1 and their corresponding labels are also indicated. All deposition experiments were performed in stationary mode, producing only a single plasma treated spot on the substrate to properly verify the deposited area as well as its homogeneity.

39

**Figure 1.** Scheme of the experimental plasma set-up: 1—mass flow controllers; 2—quartz capillary; 3—surfatron resonator; 4—microwave (MW) antenna; 5—MW coaxial cable; 6—Macor plate with precursor injection; 7—plasma torch; 8—optical fiber; 9—bubbler system in water–ice bath; 10—substrate.

**Table 1.** Summary of the selected experimental conditions. Legend: sccm, standard cubic centimeters per minute.


#### *2.3. OES*

OES was used for plasma diagnostics in an effort to reveal more information on the radiative plasma species present in the MW plasma jet. The light emitted by the discharge was integrally collected by a quartz optical cable mounted 5 cm away from the plasma output under an angle of 30◦ with respect to the plasma jet axis as shown in Figure 1. No additional optical components were used in this case, so no space-resolved OES spectra were collected. For the OES spectra acquisition, a Jobin Yvon TRIAX 550 spectrometer (Longjumeau, France) with an LN2 cooled back-illuminated CCD (1025 <sup>×</sup> 256, pixel size 26 <sup>×</sup> <sup>26</sup> <sup>μ</sup>m2) was used. Using the standard Ocean Optics DT-MINI-2-GS source (Largo, FL, USA), the system was calibrated with respect to its spectral response. These calibration results were subsequently used for measured spectra correction. The OES spectra were recorded in the range 300–850 nm using a 1200 gr/mm grating at a fixed spectrometer entrance slit of 30 μm. Four integration times (0.002, 0.01, 0.1, and 1 s) were applied for an appropriate signal/noise ratio for each studied line and band.

#### *2.4. Computational Fluid Dynamic Simulations*

To obtain information on the impact of the MW plasma jet configuration on the gas dynamics in the plasma region, computational fluid dynamic (CFD) modeling using Comsol MultiPhysics 5.3 was carried out [36]. Considering the design of the experimental set-up and the interaction of the gases with the substrate, simulations were performed under the assumption of turbulent flow [35]. Transport of concentrated species was also applied to obtain information on the distribution of the mass fraction of argon and the precursor. Capillary edge–substrate distances and carrier gas flow rates were selected as variables for computation of the mole fraction and the velocity field in the observation region. Three-dimensional geometry was used for these simulations since there was no axial symmetry in the experimental plasma set-up. Figure 2 represents the constructed geometry of the plasma set-up for the XZ- and XY-plane views. The simulations were carried out in a cylindrical surrounding region with a 150 mm radius and 130 mm height, which was assumed to be filled with ambient air, and the dimensions of all the plasma set-up parts were in accordance with the experimental set-up description of Section 2.2. Argon and a monomer-containing Ar carrier gas were injected from the main gas and side gas inlets, respectively, as shown in Figure 2b, and incompressible flows were considered. For the main gas inlet, a flow rate corresponding to 1 slm was set. For the sideway inlet, the gas flow rate was varied between 125, 250, 500, and 750 sccm. The Navier–Stokes equation and the continuity equation were solved with a fine physics controlled mesh in the volume of interest, i.e.,

$$\rho(\mathfrak{u}, \nabla)\mathfrak{u} = \nabla \cdot \left[ -p\mathcal{I} + \left( \mu + \, \mu\_T (\nabla \mathfrak{u} + (\nabla \mathfrak{u})^T) \right) \right] + \mathcal{F} \tag{1}$$

$$
\rho \nabla \cdot (\mathfrak{u}) = 0 \tag{2}
$$

where ρ is the density (kg/m3), *u* the gas velocity (m/s), *I* the identity matrix, *p* the gas pressure, μ the dynamic viscosity (Pa·s), <sup>μ</sup>*<sup>T</sup>* the turbulent dynamic viscosity (Pa·s), *<sup>F</sup>* the volume force (N/m3), and <sup>∇</sup> the differential operator.

As previously mentioned, transport of the concentrated species was also introduced, making use of the following equations, i.e.,

$$
\nabla \mathbf{j}\_i + \rho (\mathbf{u} \cdot \nabla) \omega\_i = R\_i \tag{3}
$$

$$\mathbf{N}\_{i} = \mathbf{j}\_{i} + \rho \mathbf{u} \boldsymbol{\omega}\_{i} \tag{4}$$

where j*<sup>i</sup>* kg(m2·s) is the relative mass flux component and <sup>ω</sup>*<sup>i</sup>* a mass fraction.

**Figure 2.** (**a**) XZ-plane and (**b**) XY-plane views of the MW plasma jet geometry.

#### *2.5. AFM*

The surface topography and the thickness of the deposited coatings were obtained by means of an XE-70 AFM apparatus (Park Scientific Instruments, Suwon, South Korea). In this work, the non-contact tapping mode using a highly-doped single crystal silicon cantilever with a spring constant of approximately 40 N/m was used to collect AFM images. Subsequently, using XEP software, the roughness of the surfaces was calculated. Additionally, to measure the thickness of the deposited films, a scratch was first made with a sharp blade on the surface of the cover slip glass covered by the coating. Subsequently, the scratch region was visualized using AFM and the difference in height formed on the substrate by the scratch was determined using XEP software. Each reported result is an average value calculated from three different measurements and presented with standard deviation.

#### *2.6. SEM*

After gold sputtering of the coated substrates with a JEOL JFC-1300 Auto Fine Coater (Peabody, MA, USA) for 20 s, the surface morphology of the coatings was visualized using a JEOL JSM-6010 PLUS/LV SEM device (Peabody, MA, USA) which applied an accelerating voltage of 7 kV. The magnifications used to acquire the SEM images in this work were 1000× and 30,000×.

#### *2.7. Static WCA Analysis*

The wettability of the deposited layers was measured using a Data Physics OCA 10 WCA instrument (Filderstadt, Germany). Droplets of distilled water 1 μL in volume were first placed on top of the deposited coating, after which an image of the water droplet resting on the surface was captured by a CCD video camera (Filderstadt, Germany). Afterwards, by fitting the profile of the water drop using Laplace–Young curve fitting, the WCA value was obtained. The experiments were repeated three times and the measurements were also reported as an average value with an error of less than 3.0◦.

#### *2.8. XPS*

Chemical analysis of the surface of the deposited films was assessed by performing XPS experiments. The measurements were conducted using a PHI Versaprobe II spectrometer (Kanagawa, Japan) with a monochromatic Al K<sup>α</sup> X-ray source (hν = 1486.6 eV) operating at 25 W, resulting in a beam size of 100 μm. Survey scans and high-resolution C1s spectra were acquired with 187.85 and 23.5 eV pass energy, respectively. All XPS measurements were conducted in a vacuum of at least 10−<sup>6</sup> Pa with an angle of 45◦ between the sample surface and the analyzer. By processing the survey scans using Multipak software (v 9.6) and applying a Shirley background, the elemental composition of the samples under investigation was determined. In addition, the high-resolution C1s peaks were curve-fitted by utilizing Gaussian−Lorentzian peak shapes with a 1.4 eV limitation of the full width at half maximum (FWHM) after a calibration of the energy scale was performed with respect to the hydrocarbon component of the C1s spectrum (285.0 eV).

#### **3. Results and Discussion**

#### *3.1. Discharge Characterization*

In the first stage of this study, the composition of the plasma during the coating deposition process was investigated. In an effort to obtain better insight into the quantity and quality of excited plasma species and their influence on the properties of the deposited coatings, OES was performed. The following bands and lines were observed in the obtained OES spectra: bands of OH (A→X); N2 first negative and second positive systems; and argon, carbon, hydrogen, and oxygen atomic lines. Additionally, a weak radiation of atomic hydrogen was also noticed in the OES spectra. As mentioned above, all spectra were collected using four integration times (1, 0.1, 0.01, and 0.002 s). Nevertheless, in this work, the intensity of each selected component has been presented using an integration time which was sufficiently long to provide a good intensity of signal but which avoided signal saturation at the same time. The evolutions of OH, N2 second positive system, N, Ar, Ar+, C, CH, and O as some representative lines were selected to assess information on the discharge excited species in correlation with the experimental conditions. Table 2 provides an overview of these selected components with the

wavelengths, corresponding transitions, and integration times used for their acquisition. The influence of power and carrier gas flow rate on the intensity of these lines was investigated, since these two parameters had the most significant impact on the concentration of excited species.


**Table 2.** Emission bands/lines and their corresponding transitions, wavelengths, and selected integration time.

The selected spectral lines can be divided into two different groups: one group (OH, N, N2 second positive, Ar, Ar+, and O) features the discharge itself, and those in the other (C and CH) are representatives of the injected monomers MMA and styrene. The first evident trend observed in all emission spectra was the increase in intensity with increasing MW power, which was due to the extra dissociation and fragmentation of the monomer molecules that subsequently occurred with increasing input power. In fact, higher energy delivered to the plasma resulted in an increased amount of excited species generation. The dependence of the carbon line emission intensity on the carrier gas flow rate at different input powers is presented in Figure 3 for the two monomers under study (styrene and MMA). Based on these graphs, it can be concluded that precursor fragmentation was almost twice as significant when using MMA as a monomer instead of styrene. This difference can be explained by the different levels of dissociation energy required for breaking the bonds in each monomer molecule. Immediately after introducing the monomer to the system, the intensity of the carbon line also increased for both monomers under study. The intensity line sharply increased with increasing carrier gas flow rate until it reached its maximum value at a carrier gas flow rate of 250 sccm. Further increasing the monomer input, however, resulted in a decrease in the intensity of the atomic carbon line, demonstrating that a higher monomer concentration present at the same power level results in a decrease in available fragmentation energy per monomer molecule. Moreover, for a higher carrier gas flow rate, the interaction time between monomer molecules and plasma reactive species was shorter, which in turn decreased the concentration of the carbon species. The same trend with maximum intensity at a 250 sccm carrier gas flow rate was also observed for the CH compound, as can be seen in Supplementary Materials Figure S1.

The optical emission of the Ar and Ar<sup>+</sup> lines is proportional to the concentration of reactive species in the argon discharge. Thus, the dependency of these line intensities on the carrier gas flow rate and applied power can provide insight into the processes occurring during thin film deposition. The evolution of the Ar and Ar<sup>+</sup> line emission intensity as a function of carrier gas flow rate at different discharges powers is presented in Figure 4a,b for styrene and MMA, respectively. A comparison between Figure 4a,b suggests that for a higher applied power, the intensity of Ar and Ar<sup>+</sup> were almost independent of the type of monomer which was present in the discharge, while for lower power, an admixture of MMA resulted in an enhanced intensity of Ar and Ar<sup>+</sup> lines in comparison with styrene. In addition, the signal from these two argon lines increased with increasing carrier gas flow rate and again reached its maximal level at a carrier gas flow rate of 250 sccm for both monomers under study. The growing trend can be explained by the higher amount of overall argon gas in the system which flowed with a higher speed. At a carrier gas flow rate of above 250 sccm, a steady state of the intensity of the Ar and Ar<sup>+</sup> lines appeared, which can be seen to be related to the air admixing and the presence of a higher concentration of monomer molecules in the discharge. The

same trend of intensity profiles was also observed for the N2 second positive, N, OH, and O bands (see Figures S2–S5, Supplementary Materials). The similarity in the optical emission intensity between all these mentioned bands and the argon lines thus indicates that the processes standing behind the excitation were strongly interconnected.

**Figure 3.** Dependence of the C line emission intensity on the carrier gas flow rate at different powers for the (**a**) styrene and (**b**) methyl methacrylate (MMA) monomers.

**Figure 4.** Dependence of the Ar and Ar<sup>+</sup> line emission intensities on carrier gas flow rate at different applied powers for the (**a**) styrene and (**b**) MMA monomers.

Making use of the integral intensity of a few selected argon lines, namely the lines at 603.2, 675.3, 687.1, and 714.7 nm, the excitation temperature of argon was also calculated using corresponding constants from the NIST (National Institute of Standards and Technology) database [37]. The calculations resulted in excitation temperatures in the range 4300–5000 K under all experimental conditions and for both monomers. In general, no dependence on the carrier gas flow rate was observed for the excitation temperature. However, the excitation temperature was slightly higher with increasing power. In the next step, a Boltzmann plot from the N2 second positive 0-2, 1-3, and 2-4 band heads was used to determine the vibrational temperature. This temperature was found to be dependent on discharge power and monomer type and was rather stable when changing the Ar carrier gas flow rate. For the MMA precursor, the vibrational temperature varied between 2100 and 2900 K depending on the applied power while this temperature range was slightly lower for styrene, being 2000–2800 K. Furthermore, according to the Boltzmann distribution, using the N2 second positive 0-2 band with J = 27–32, the rotational temperatures were also calculated [38]. For this calculated rotational temperature, no significant impact of power, monomer type, and carrier gas flow rate variation on the obtained value was observed. The range of rotational temperature variation for both monomers was

found to be 1250–2500 K, with a mean uncertainty of 250 K. Additionally, the rotational temperature calculated from the OH radical lowest lines (details in [20]) was 550–700 K, with a mean uncertainty of 70 K.

#### *3.2. CFD Simulations*

Additional to the plasma characterization, 3D numerical simulations of fluid dynamics were also carried out in this study to obtain a basic perception of how the mixing of the main Ar and the Ar carrier gas stream occurred. Moreover, information on the distribution of the monomer in the volume close to the substrate was also able assist in evaluating the thin film growth during the deposition process. For example, it has already been shown that the gas velocity and mass fraction can be correlated to the deposition pattern on a substrate [39] and the plasma behavior [35]. In the latter study it was also revealed that placing an additional plate at the edge of the plasma jet capillary intensified the gas purity near the surface when working with a small gap distance to the substrate [35]. In this study, the experimental set-up was quite similar to that used in [35], as the additional disc for the purpose of monomer introduction also played the role of an additional plate and thus constrained the gas in a small volume, which in turn influenced the gas diffusion and propagation near the substrate. All computations in this work were only performed for the styrene monomer admixture, since its deposition pattern on the substrate differed from a simple circular shape. This was not the case when using MMA as monomer, as will be shown in the following sections.

The first stage of the simulations was focused on defining the velocity distribution in the MW plasma jet system. In this step, the main gas flow rate was fixed at 1 slm and the distance between the capillary and the substrate was set to 5 mm. Computations were performed for different monomer-containing gas flow rates (125, 250, 500, and 750 sccm), and Figure 5 demonstrates the gas velocity for these different gas flow rates in the XZ-plane. Figure 5a shows that for the smallest value of the carrier gas flow, there was no disturbance to the main gas stream, as a higher velocity was present in the larger channel compared to the sideway gas inlet. Figure 5b also shows that when the additional gas reached a flow rate of 250 sccm, the velocity of the two gas streams became comparable at their meeting point. Moreover, when using even higher carrier gas flow rates (Figure 5c,d), the impact of the sideway gas inlet became more pronounced. Under these experimental conditions, the side gas flow could not easily deviate from its direction and could not mix well with the main gas at their meeting point. The higher velocity of the sideway gas pushed the monomer-containing gas and the main gas flow on the side of the capillary opposite to the sideway gas inlet, as shown in Figure 5c,d. Changes in the velocity of the gas near the substrate can be seen as well in Figure 5. The homogeneous velocity distribution of the gas in the case of small side gas flow rates changed to an uneven distribution when injecting sideway gas with higher flow rates. It is also worth mentioning that the diameters of the two gas inlets were different (3.1 mm for the main gas stream and 2 mm for the sideway gas inlet), which influenced the velocity magnitude. The high velocity of the main argon gas stream in the center could also prove that there should be a negligible amount of deposition in the center of the substrate, especially when using low side gas flow rates.

In a next step, the mass fraction of the monomer precursor and the way the monomer propagated on the surface were simulated. The results are shown in Figure 6 for the XY-plane (top view) and in Figure 7 for the XZ-plane of the used 3D model. The XY-plane crossed the volume at the position where the substrate surface was located during the experiments. The shown series of graphs indicates that for the smallest carrier gas flow rate (125 sccm), only a small amount of monomer was present on the substrate surface and that the monomer completely diverged to the right-hand side of the plasma set-up. An increase in the monomer-containing gas flow rate to 250 sccm intensified the presence of the monomer at the right side of the set-up with only the presence of a negligible amount of monomer on the other side of the substrate. On the other hand, when the flow rate of the side gas was further increased to 500 sccm, a small propagation of the monomer to the left side of the substrate was observed. Moreover, a 750 sccm side gas flow rate even led to the monomer spreading over almost the complete area of the substrate. In the XZ-plane of the 3D model, depicted in Figure 7, the majority of the monomer also deviated to the right-hand side below the opening of the Macor disc on the substrate surface when the carrier gas flow rate was small (Figure 7a,b). For these studied conditions, the velocity of the main gas stream was much higher than the velocity of the gas stream coming from the smaller side channel. Consequently, the monomer only slightly mixed with the main stream of the system and remained on the same side when the gas outflowed the capillary. On the other hand, in the cases where carrier gas flow rates of 500 and 750 sccm were used (Figure 7c,d), the mixing process of the two streams was significantly improved and the monomer was more directed to the center, since it was entering the channel with a higher velocity. Moreover, at the highest flow of the side gas, the volume between the Macor disc and the substrate was almost entirely filled with monomer, which was not the case for the other analyzed carrier gas flow rates. It can therefore be concluded that increasing the carrier gas flow rate resulted in a more uniform monomer dispersion on the substrate surface based on the mass fraction of the precursor in the volume between the plasma jet and the substrate. These results of the monomer propagation region shape can thus provide a preliminary prediction of the footprint of the deposited layer on the substrate.

**Figure 5.** Velocity distribution (m/s) in the XZ-plane at a 5 mm gap size between the capillary and the substrate for a sideway gas flow rate of (**a**) 125 sccm, (**b**) 250 sccm, (**c**) 500 sccm, and (**d**) 750 sccm.

**Figure 6.** Top view (XY-plane) of the monomer mass fraction distribution at 5 mm gap size for (**a**) 125 sccm, (**b**) 250 sccm, (**c**) 500 sccm, and (**d**) 750 sccm side gas flow rates.

**Figure 7.** XZ-plane cross-sections of the monomer mass fraction distribution at 5 mm gap size for (**a**) 125 sccm, (**b**) 250 sccm, (**c**) 500 sccm, and (**d**) 750 sccm side gas flow rates.

The size of the volume between the Macor disc and the sample was regulated by the gap size between these two planes. Consequently, the working distance also had a significant influence on gas propagation at the capillary outlet. Figure 8 shows the influence of the gap size on the monomer distribution for distances of 3, 5, and 7 mm between the bottom of the Macor disc and the substrate plane. In these simulations, the sideway gas flow rate was fixed at 250 sccm. For a smaller distance (Figure 8a), the volume of propagation was too small to provide a sufficient distance for efficient mixing of the two gas streams. As a result, the pattern of the monomer fraction was in this case mainly controlled by the main gas flow. A similar behavior was also observed when using a 5 mm distance, yet a more uniform gas propagation on the right side of the sample was present due to a longer available distance for mixing of the two gas streams. Finally, for a 7 mm distance between the disc and the substrate (Figure 8c), the mass fraction of the monomer was more evenly distributed on the surface of the substrate, which might in turn have resulted in a better homogeneity of the deposited coating.

**Figure 8.** Top view (XY-plane) of the monomer mass fraction distribution for (**a**) 3 mm, (**b**) 5 mm, and (**c**) 7 mm gap sizes. The monomer-containing gas flow rate was fixed at 250 sccm.

#### *3.3. Physical and Chemical Surface Characterization*

#### 3.3.1. Coating Appearance and Homogeneity

In the next step of this study the physical and chemical properties of the deposited coatings were investigated. For this purpose, the MW plasma jet was used for thin film deposition of MMA and styrene coatings. Initially, both monomers were deposited on a silicon wafer, making use of the main conditions (power: 10 W, monomer-containing gas flow rate: 250 sccm, deposition time: 3 min, and gap size: 5 mm (see Table 1)) to get an idea of the deposited thin film pattern on a flat substrate. The obtained plasma-deposited thin films were studied in terms of surface morphology and topography by utilizing SEM and AFM, respectively, and the obtained results have been depicted in Figure 9. As can be seen in this figure, each coated silicon wafer had a unique appearance depending on the used monomer and the way it was mixed into the MW plasma jet effluent. The diameter of the coated area for the plasma-deposited styrene was approximately 20 ± 5 mm, while for MMA, the diameter was reduced to 3.0 ± 0.5 mm. However, in the case of styrene, the deposition pattern on the substrate was not entirely symmetrical. This result can be explained by the non-homogeneous mixing of the monomer into the main stream of the plasma, as was shown in the previous section. On the other hand, when MMA was added to the plasma, a coating located near the center of the substrate was deposited, as demonstrated in Figure 9b. Figure 9 thus clearly shows that the styrene coating had larger dimensions compared to the MMA-based film deposited under the same experimental conditions. A possible explanation for this dissimilarity may be the difference in deposition rate of the two monomers under study. For instance, organic compounds containing double bonds and aromatic rings are known to have a higher deposition rate in comparison to branched-chain or non-aromatic structures [40]. Hence, styrene is expected to show a higher deposition rate than MMA, which can in turn result in a larger deposition area.

**Figure 9.** Overview of the appearance, surface morphology, and surface topography of thin films deposited with the MW plasma jet by injection of the (**a**) styrene and (**b**) MMA monomers. The observed area of the 3D atomic force microscopy (AFM) image is 30 <sup>×</sup> <sup>30</sup> <sup>μ</sup>m2.

On the styrene-based coating area, five different zones have been determined which all showed a different surface topography and morphology. Based on Figure 9a, the following zones were selected: zone C—directly under the MW plasma jet capillary where the coating appeared in the form of dispersed droplets with round shapes and sizes reaching up to 1 μm; zone B—located above zone C, consisting of a very smooth and thick coating; zone D—located below zone C where no deposition at all could be observed; and zone A and zone E—located at the edges of the coated area. In these last two regions, the deposits were characterized by a very thin non-uniform coating that contained plenty of vacant spots. The location of the MMA coating, as shown in Figure 9b, matched the position of zone B of the styrene-based sample, suggesting similarity in gas dynamics for both monomers as the same plasma jet device was used. Indeed, a higher amount of deposited precursor was situated above the center of the substrate as the monomer gas injection point was located in this zone. This not-centered position of the deposited coating can be explained by the suppressive behavior of the main gas stream towards the injected monomer. Since the monomer is injected in the region where the plasma jet is fully developed, well-structured, and close to the outlet, it does not have enough residence time as it propagates towards the core of the plume. Thus, the majority of the monomer is not deposited in the central area but in the zone shifted towards the monomer injection side. Figure 9b also shows that the coating morphology was different when MMA instead of styrene was used as a monomer, which could be attributed to the lighter weight of MMA molecules compared to styrene. The OES results suggest that the plasma jet with injected MMA contained a higher amount of excited species which could have interacted with the monomer molecules. This could in turn have led to nucleation in the gas phase, which is followed by condensation and finally deposition in the form of droplets on the surface. This effect was also valid for styrene but was less pronounced, as shown in Figure 9a.

#### 3.3.2. Coating Morphology

In this part of the study, the morphology of the coatings was examined using SEM imaging. As mentioned above, the majority of the coated sample surfaces contained droplets distributed all across of the deposit. Figure S8 (Supplementary Materials) depicts a cross-sectional view of the coatings on the glass substrate under the main conditions. The formation of droplets and clusters on the sample surface is a sign that monomer plasma polymerization actually occurs during the coating deposition [25,41,42]. Based on the used experimental conditions, the size and distribution of these deposited droplets were found to fluctuate. Figure 10 shows the SEM images of both styrene (zone C) and MMA samples prepared with different carrier gas flow rates. Generally, deposited styrene in the center contains droplets with a smaller size compared to the MMA monomer for all experimental conditions. A common trend can also be noticed for the SEM images presented in Figure 10: increasing the monomer-containing gas flow rate caused an increase in the number of droplets formed on the surface of the coatings. Additionally, by adding monomer to the system the size and shape of the droplets also varied. For example, when using 750 sccm MMA, the large amount of injected monomer resulted in a surface coverage by a high amount of droplets which appeared as a smooth coating. The influence of other working parameters (time, power, and distance to substrate) on the surface morphology can be observed in Supplementary Materials Figures S6 and S7 for MMA and styrene, respectively. A short deposition time (1 min) and lower carrier gas flow rate resulted in coatings possessing a smaller amount of droplets, while higher values of these parameters produced deposits with a higher amount of droplets across the investigated area for both monomers. In addition, the distance between the MW plasma jet disc and the substrate influenced the size of the droplets: increasing this distance to 7 mm made the size of the deposited droplets larger. This can be explained by the gas dynamics in the volume between the Macor disc and the substrate. A smaller distance can provoke vortex creation at the center [14], which can affect the location of droplet creation on the substrate surface, while for a larger distance this effect is less prevalent, resulting in deposits with larger-sized droplets. To prove that a plasma-polymerized cross-linked coating was deposited on the substrates in this work, the coatings were submerged in water for a short duration. This was tested for the main condition deposits and Figure S9 (Supplementary Materials) confirms that the coating was still present on the substrate after immersion in water for 24 h. This result clearly evidences that a plasma-polymerized cross-linked coating was deposited in this work [43,44].

#### 3.3.3. Coating Thickness

As previously mentioned, to measure the thickness of the deposited film, a scratch was first made with a sharp blade along the zones of the samples that were defined on the surface of the coating deposited on the cover slip glass. Subsequently, the scratch region was visualized using AFM and the difference in height formed on the substrate by the scratch was determined. Within this context, it is important to mention that during the making of the scratches, the cover slip glasses remained untouched, making the thickness measurement more reliable. As previously shown in Figure 9, it had already been observed that a thick homogeneous coating was obtained in zone B of the plasma-deposited styrene sample. As a result, in the case of styrene, the thickness of the films was measured in zone B. Coating thickness measurements were, however, not performed for samples prepared using MMA as a monomer, as in this case the coating was only spread across the substrate in the form of droplets. Consequently, Table 3 only contains the coating thickness results for styrene samples prepared under different experimental conditions.

**Table 3.** Thickness of the styrene-based coating under different experimental conditions.


**Figure 10.** SEM images of the central zone of styrene (zone C) and MMA coatings deposited with different carrier gas flow rates.

These results demonstrate that the lowest measured thickness corresponded to the lowest deposition time (1 min), and, conversely, that the highest deposition time resulted in the thickest coating. Another parameter that had a direct influence on the coating thickness was the carrier gas flow rate. As is reported in Table 3, the smallest Ar flow-carrying styrene (125 sccm) produced a coating with a 110 nm thickness after a deposition time of 3 min. When a higher Ar carrier gas flow rate of 250 sccm was used, the coating thickness reached its maximum level of 229 nm after the same deposition time. Interestingly, any further increase of the carrier gas flow rate did not result in a thicker coating but

resulted in a more homogenous film deposition over the substrate area as other measurements along the scratch in this zone showed only a small deviation from the reported coating thickness. Based on the simulation results reported in Section 3.2, it was concluded that at the highest carrier gas flow rate (750 sccm), the carrier gas was injected into the system with a considerably high velocity, which could in turn spread the monomer all over the substrate surface due to efficient mixing with the main plasma jet gas stream. Besides the gas propagation behavior, which only predicted a better uniformity of the coating at the highest carrier gas flow rate, the available energy for monomer fragmentation should also be taken into account to explain why the coating thickness did not increase when increasing the Ar carrier gas flow rate to 500 and 750 sccm. Since the applied power remained the same while the carrier gas flow rate was varied, the average energy per monomer molecule reduced with increasing monomer injection flow. As a result, the efficiency of precursor fragmentation was also less pronounced at higher sideway gas flow rates, which is in agreement with the OES results shown in this work. Although one would expect a thicker coating with increasing carrier gas flow rate as a higher amount of monomer was injected in the system, this increase in monomer amount was in this study counteracted by the lower amount of monomer fragmentation, ultimately leading to the deposition of a thinner coating.

Table 3 also reveals that the distance between the disc and the substrate also influenced the thickness of the coating: a higher distance to the substrate yielded a thicker and more uniform coating, while a smaller gap size provided a wider coated area with a lower coating thickness and less uniformity. The latter conclusion concerning coated area and coating uniformity was drawn based on thickness measurements which were performed along the scratch. These results can again be well correlated with the simulations performed in this work, showing the effect of distance on the monomer distribution. Moreover, at a lower distance, a developed turbulent flow of main and carrier gas can occur, which can in turn influence the size and homogeneity of the coated area. Finally, Table 3 also shows that the coating thickness was also affected by the applied power. An increase in power to 10 W resulted in a higher coating thickness, resulting from more monomer fragmentation at elevated power. However, at a power of 13 W, the coating thickness was not increased further, which could be explained by the fact that in this case monomer fragmentation was too pronounced to efficiently deposit a coating.

#### 3.3.4. Coating Roughness

Another critical coating parameter that can be measured by AFM is surface roughness, and the obtained roughness results are presented in Table 4. Generally, for styrene samples, the highest surface roughness was observed in the center of the deposit (zone C), as in this area the monomer was mostly deposited in the form of dispersed droplets with varying sizes. This roughness, however, decreased in zone B in comparison to central zone C, as indicated by the results shown in Table 4. Noting that the roughness for the pristine glass is 5 nm, zone B did contain a coating, and this coating was quite smooth, as the roughness varied between 14 ± 2 and 32 ± 6 nm, depending on the experimental conditions. These results are thus consistent with the AFM image of zone B, shown in Figure 9a. After this smooth coating region, an area covered by a non-homogeneous coating with different vacant sites was observed to follow (zone A). In this region, the roughness was found to be rather similar to or slightly higher than the roughness values obtained in zone B, depending on the applied experimental conditions. Table 4 also reveals the influence of deposition time on the surface roughness: for both monomers, an increase in deposition time resulted in a higher surface roughness. This finding can be attributed to a higher number of accumulations of the monomer on the surface at random locations when using longer treatment duration. Consequently, this monomer distribution in the form of droplets resulted in an increase in coating surface roughness. Another experimental parameter that was found to influence the surface roughness was the applied power: for both monomers, an increase in power resulted in a rougher deposit. This can be explained as follows: at a higher power, a higher level of monomer fragment reactions occurred in the gas phase instead of on the gas–substrate interface, thereby leading to a rougher coating surface. In contrast to deposition time and power, the influence of carrier gas flow rate on the surface roughness was dependent on the type of injected precursor (MMA or styrene). For MMA, the highest surface roughness was obtained when using the lowest carrier gas flow rate, and this roughness decreased with increasing flow of monomer into the system. As stated above, with an increase in the amount of injected monomer to the system the droplets on the substrate recombined with each other and decreased the surface roughness and finally completely covered the surface at the higher flow rate, producing a smooth coating. In the case of styrene, the smallest carrier gas flow rate of 125 sccm produced films with a comparatively small roughness in all zones. These roughness values in the center of the sample (zone C) increased when a 250 sccm monomer-containing gas flow rate was used, and then again decreased when this gas flow rate was further increased to 500 and 750 sccm. This behavior can be explained by a combination of monomer distribution and gas velocity during the deposition process. Zone C is just below the plasma jet capillary, and, thus, with a 125 sccm carrier gas flow rate, the monomer was not well mixed with the main stream, leaving this area almost without a coating (just a few deposited droplets). A higher amount of droplets was able to be observed for the next value of Ar carrier gas, since in this case more monomer fragments were delivered to the surface by the faster gas stream. Lastly, a smooth layer in zone C was observed when deposition was performed at higher flow rates (500 and 750 sccm) due to good monomer mixing with the main gas stream and its uniform distribution under the plasma jet capillary, as was shown in Section 3.2. In zones A and B, which were located at the side where the monomer was injected, the values of surface roughness increased with increasing monomer-containing gas flow rate. These findings can be explained by the simulation results obtained earlier. It was shown that at a lower flow rate the monomer was not well mixed with the main stream and mostly remained on one side of the plasma jet, while higher values of carrier gas flow rate caused direct injection of the monomer into the core of the MW jet. Thus, the nucleation of styrene near zones A and B proceeded at different radial positions of the plasma jet: near its edge or in the core for smaller and higher side gas flow rates, respectively. This different behavior resulted in coatings with different roughness values at these positions on the substrate, depending on the amount of injected monomer. Finally, the distance between the MW plasma jet and the sample was found to have an insignificant influence on surface roughness.

**Table 4.** Surface roughness of styrene and MMA deposited thin films under different experimental conditions.


#### 3.3.5. Coating Wettability

In this part of the study, the wettability of the deposits prepared in this work was determined utilizing WCA analysis. The measurements were again performed on the styrene and MMA samples in the zones that were introduced earlier in this work. The wettability of an uncoated cover slip glass was measured to be 79◦ ± 2◦. After plasma deposition, the WCA values of the coated samples significantly decreased, suggesting that the coated layers were hydrophilic. Summarized results of WCA measurements for both monomers and under different experimental conditions are presented in Table 5. These results indicate that shorter deposition times and smaller carrier gas flow rates resulted in coatings with higher hydrophilicity for both precursors. Consequently, a higher flow rate of the monomer-containing gas produced deposits with higher WCA values. This latter behavior can be explained by a reduction in monomer molecule fragmentation, and, therefore, an incorporation of more non-oxidized long-chain monomers on the surface with increasing carrier gas flow rate. Furthermore, Table 5 also shows that an increase in applied power resulted in more hydrophobic coatings. A possible explanation for this can be that there was lower fragmentation with lower power. In the case of styrene, the highest wettability was for all experimental conditions observed in zone C, the location precisely under the plasma exposure point, which might be due to the fact that this point had the best

interactions with the plasma jet during the deposition process. However, while radially proceeding to the edge of the sample, in the direction of zone A, a reduction in coating wettability was observed for all experimental conditions under which styrene coatings were deposited. Only a small difference in WCA was observed for different distances between the MW plasma jet and the substrate, which can be explained by variations in the gas flow dynamics in the volume above the sample.


**Table 5.** Water contact angle (WCA) angle values for styrene and MMA coatings prepared under different experimental conditions.

#### 3.3.6. Coating Chemical Composition

Chemical surface analysis is one of the most valuable characterization tools for deposited coatings, as it provides information on thin film surface composition. The elemental compositions of the coatings prepared under various experimental conditions were determined by XPS in addition to the concentration of the carbon-containing chemical bonds. Similarly to WCA, the coatings were again analyzed in the same zones as where all other measurements were performed. From XPS survey spectra it was found that the coatings produced from styrene only consisted of carbon and oxygen, except for the central point (zone C), at which a small amount of nitrogen was also detected. It is well known that APPJ treatment can induce the incorporation of oxygen by post-treatment exposure of the sample to the atmosphere or by reactions with oxygen which are present in the discharge due to gas impurities and ambient air admixture [45,46]. OES measurements demonstrated that in this work a considerable amount of oxygen and nitrogen species were present in the plasma jet during the thin film deposition when using styrene as a monomer. This can explain why in case of styrene, which contains no oxygen in its chemical structure, it was possible to detect oxygen and nitrogen in the deposited coatings. In the case of MMA, the surface of the thin films only consisted of carbon and oxygen, regardless of the location or the applied experimental conditions, which is in agreement with the chemical structure of the original monomer. Based on the elemental composition results, O/C ratios were calculated for MMA and all regions of the styrene samples, and the obtained values have been presented in Supplementary Materials Table S1. These results were found to be in good agreement with the earlier obtained WCA results, since a higher oxygen incorporation (higher O/C ratio) corresponded to a more hydrophilic region. Moreover, for the styrene samples, the O/C ratio followed the same trend across the surface as was noticed before for the coating wettability properties. A higher amount of oxygen was incorporated in the central area (zone C), and each subsequent area towards the edge of the coating contained less oxygen.

Deconvolution of high-resolution C1s spectra was also performed to obtain more detailed information on the chemical structure of the deposited thin films. The C1s high-resolution spectra for deposited styrene coatings were fitted with five curves, namely, (1) aromatic C–C/C–H bonds at 284.8 eV; (2) aliphatic C–C/C–H bonds at 285 eV; (3) the oxygen-containing group C–O at a binding energy of 286.6 eV; (4) O–C–O/C=O groups at a binding energy of 288.1 eV; and (5) a shake-up peak at 291.5. For MMA samples, curve-fitting of C–C/C–H, C–COO, C–O, O–C–O/C=O, and O–C=O were used at binding energies of 284.8, 285.4, 286.6, 287.9, and 289.1 eV, respectively. Figure S10 (Supplementary Materials) shows a typical deconvolution of the detailed C1s spectra for styrene and MMA deposits with the indication of the reference peaks ascribed to conventional PS (polystyrene) and PMMA (poly(methyl methacrylate)) as well as new peaks induced during plasma polymerization [47]. The discrimination of the standard peaks attributed to the mentioned polymers is evidence that the plasma polymerization process was happening during the coating deposition with MW APPJ for both monomers. For the MMA coatings, the appearance of the peak at 287.9 eV could have been due to the scission of C–OCH3 bonds during plasma exposure, as was reported in the literature [48]. In the case of styrene, the presence of two extra peaks attributed to oxygen-containing groups suggests the breakage of C–C/C–H bonds, followed by recombination with oxygen from the surrounding environment [49]. The presence of the minor shake-up bond revealed that by using the MW plasma jet for thin film deposition, the benzene ring of the styrene monomer was at least partly preserved in the deposits. Figure 11a,b show the C–O/C–C ratio as a function of carrier gas flow rate for styrene and MMA, respectively. From these figures, it is clear that for both monomers the amount of C-O functional groups was highly dependent on the rate of monomer injection into the system. In all regions of the styrene coating and in case of the MMA coating, the smallest flow rate of the carrier gas (125 sccm) resulted in the highest values of C–O/C–C ratio. Increasing the carrier gas flow rate up to 500 sccm led to the incorporation of a higher amount of carbon in the form of C–C/C–H bonds and, consequently, the C–O/C–C ratio decreased. Further increasing the carrier gas flow rate to 750 sccm, on the other hand, did not significantly affect the C–O/C–C ratio anymore for both monomers under study. This trend can be explained by the concentration of the monomer in the discharge during its fragmentation: at a lower carrier gas flow rate, fragmented monomer molecules mostly interacted with oxygen molecules present in the plasma rather than with other monomer fragments which occurred at higher rates of monomer injection. Hence, the C–O/C–C ratio was lower in the case of higher carrier gas flow rate due to more pronounced recombination processes between the monomer fragments. The influence of the applied power, the distance between the plasma jet disc and substrate, and the deposition time on the coating chemical composition was also investigated and the results have been reported in Figure S11 (Supplementary Materials) together with the carrier gas flow rate results for both styrene and MMA. This figure reveals that the deposition time and the jet-substrate distance had a mostly insignificant impact on the C–O/C–C ratio: generally speaking, all these samples only varied in a small range of C–O/C–C ratio with varying time and distance. By contrast, the discharge power was found to have an impact on the C–O/C–C ratio: in the case of both monomers, a higher C–O/C–C ratio was obtained for a lower applied power. This result could be explained by the presence of other oxygen-containing functional groups such as C=O and O-C=O at higher applied powers. Finally, Figure 11 and Figure S11(Supplementary Materials) also reveal that for all applied experimental conditions, zones B and C had the highest C–O/C–C ratios, suggesting that the C–O functionalities were mainly present in the region close to the MW plasma jet.

**Figure 11.** Influence of carrier gas flow rate on C-O/C-C ratio for (**a**) styrene and (**b**) MMA samples.

#### **4. Conclusions**

In this study, the feasibility of MMA and styrene deposition by means of a MW APPJ has been investigated. For the first time, this type of plasma source has been applied to deposit thin layers under atmospheric pressure. Using different techniques, the effect of process parameters such as applied power, carrier gas flow rate, distance from capillary to the substrate, and treatment time on the deposition efficiency has been studied. Firstly, using OES, a MW plasma jet was characterized in terms of excited species. The results demonstrated that a higher amount of reactive species was present in the plasma containing MMA as the monomer compared to the plasma jet containing styrene. It was also observed that the Ar and C bands had a well-defined maximum of OE spectra intensity which was reached when adding 250 sccm of the monomer-containing gas into the system. It can therefore be concluded that the optimal fragmentation of the introduced monomers and argon atom excitation occurred under mid-experimental conditions. CFD simulations were also employed to determine the pattern of precursor propagation in the volume close to the substrate surface. The obtained results of the CFD simulations displayed a remarkable resemblance between the pattern obtained in the experimental stage and the CFD simulations. It was shown that the injection of the monomer from one side caused non-uniformity of the coating due to the non-homogeneous precursor distribution in the examined system. Additionally, an increase in carrier gas flow rate also influenced the pattern of the deposited layer on the substrate and enforced the monomer to propagate further into the main gas stream, thereby providing better mixing. As a conclusion, CFD simulations thus predicted the distribution of the coating on the substrate during the experiments. Coating thickness, roughness, and topography were evaluated using AFM measurements and SEM imaging. The results demonstrated the appearance of droplets in the central area of the deposited styrene coatings while the areas further away from the center showed almost empty regions in one direction and a comparatively thick coating layer in the other direction due to monomer distribution in the volume close to the substrate, as was demonstrated by Comsol simulations. Finally, the edges of the coated area contained a thin coating with vacant spots across their surface. In the case of MMA monomer deposition, the formation of droplets was only observed in the center within a small deposition area. Finally, WCA and XPS measurements provided information about the surface chemical properties of the deposits. Good agreement between surface wettability and oxygen incorporation was demonstrated. Higher hydrophilicity and higher O/C ratios were observed in zones closer to the center where better contact with the MW plasma jet was established. Based on C–O/C–C ratios calculated for all experimental conditions for both monomers it was concluded that the highest ratios were obtained at low carrier gas flow rates due to the effective fragmentation and oxidation. The obtained results with support from the literature allow for us to conclude that the styrene and MMA monomers were plasma-polymerized by the MW APPJ used. In this work, a well-established correlation between the plasma diagnostic results and coating surface properties was achieved. A high level of monomer fragmentation observed with OES, together with the non-uniform distribution of the monomer presented by CFD simulations, was shown to be a reliable indicator of coating quality. In conclusion, it can be asserted that this study gives useful insights into the properties of deposited films using an atmospheric pressure MW plasma jet with different working parameters. The carrier gas flow rate was found to have the most significant impact on the properties of the thin deposited films due to its effect on the gas flow dynamics in the system.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/2/354/s1.

**Author Contributions:** M.N. wrote the manuscript, conducted most of the experimental work, collected and interpreted the data, and performed the simulations; F.K. also carried out deposition experiments, prepared samples, and performed OES measurements and analysis; Y.O. supervised the work, performed data interpretation, and conducted XPS measurements; Z.K. helped conduct experimental work and additional analysis; and R.M. and N.D.G. supervised and financed the work, aided in the interpretation of the results, and worked on the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by a Starting Grant of the Special Research Fund of Ghent University (project number 01N00516).

**Acknowledgments:** The authors acknowledge the support from the Special Research Fund of Ghent University. **Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
